INTRODUCCIÓN

Esta asignatura, al igual que todas las de la Especialidad: Lengua Extranjera (Inglés), debe
contribuir a la formación de los estudiantes normalistas como futuros maestros de inglés en las
escuelas secundarias y, por lo tanto, su orientación es práctica más que teórica y pedagógica más
que académica.

El curso se realiza en español y en inglés; sin embargo, las clases se darán principalmente en inglés,
fortaleciendo así el dominio del idioma por parte de los estudiantes. En la comparación entre el
español y el inglés se destacarán semejanzas y diferencias que tengan más relevancia tanto en el
aprendizaje del inglés por parte de personas cuya primera lengua es el español, como en su
enseñanza a esas mismas personas. Además, se examinará el papel de los “errores de
interferencia”; es decir, aquellos que parecen originarse en la transferencia de formas o hábitos de
la lengua materna, pero en el contexto de la amplia gama de errores que se “cometen” durante el
largo proceso de aprender el inglés como lengua extranjera, incluyendo los “errores de adquisición”
que son universales, sin importar la lengua materna de quien aprende.

Algunas actividades propuestas en el programa pueden relacionarse explícitamente con la

preparación de clases de inglés; lo que contribuye tanto a la formación pedagógica como lingüística
de los estudiantes normalistas. Asimismo y considerando la importancia, en la enseñanza, de la
planeación de clases y la anticipación de posibles problemas y sus soluciones, se propone ampliar
los contenidos de la asignatura más allá de la gramática, también es importante que el maestro
anticipe posibles problemas léxicos, fonológicos y gramaticales.

Al aplicar el análisis comparativo a la enseñanza del inglés, es necesario recordar que existen varias
teorías acerca de los errores que se cometen al aprender una lengua, con cuerpos de investigación
bastante amplios. La teoría que más se relaciona con el análisis comparativo entre la lengua meta y
la materna es el conductismo (behaviorism), que considera que los errores tienen su origen, sobre
todo, en la interferencia de la lengua materna. Ejemplo de un probable error de interferencia es
cuando se dice: I have 20 years, en lugar de I am 20 (years old), que indica Yo tengo 20 años. El
trabajo de lingüística comparativa de Stockwell, Bowen y Martin (1965), entre otros, se inspiraba en
la teoría conductista.

Sin embargo, durante la última parte del siglo XX dominó otra teoría, la cognoscitiva (mentalism),
que considera los errores como producto natural del proceso de adquisición de la lengua y que son
universales, sin importar la lengua materna. Un ejemplo de esto es la siguiente clasificación: He like
apples, I didn’t went home, y This is more bigger; en vez de: He likes apples, I didn’t go home, y
This is bigger. Errores de adquisición universales que inclusive los niños británicos y
norteamericanos cometen al adquirir el inglés como lengua materna. Este enfoque, el “análisis de
errores”, empezó a rebasar al análisis comparativo en la teoría general del aprendizaje de las
lenguas extranjeras a partir de 1970 (véase, por ejemplo, Richards, 1974).

Actualmente, la postura de los expertos menos extremistas se sigue inclinando hacia la teoría
cognoscitiva; es decir, por la clasificación de la mayoría de los errores como resultado del proceso
de adquisición (y sobre todo de la simplificación y la sobre generalización). Sin embargo, reconocen
que muchos errores pueden tener, cuando menos, un elemento de interferencia de formas y usos
específicos de la lengua materna. Asimismo, la interferencia de la lengua materna se evidencia más
cuando todo un grupo de estudiantes tiene la misma lengua materna, que en grupos de nacionalidad
y lengua materna mixtas. El peligro de la consolidación de errores de interferencia en grupos
monolingües también es probablemente mayor en cuestiones de pronunciación.

La comparación entre el español (lengua materna) y el inglés (lengua meta) puede contribuir de
manera significativa en la planeación y enseñanza de clases de inglés. Pero debe cuidarse que, para
los estudiantes normalistas, el conductismo y el análisis comparativo no se conviertan en las
principales claves psicológicas y lingüísticas para la enseñanza y aprendizaje del idioma, ya que son
procesos sumamente complejos sobre todo en el aprendizaje realmente efectivo.

1
 PROPÓSITOS GENERALES

Los propósitos generales de la asignatura Elementos Básicos de Gramática Comparada Inglés-

Español son:

1. Continuar desarrollando el dominio integral del inglés.

2. Concienciar a los estudiantes de las semejanzas y diferencias más significativas entre el español
y el inglés.

3. Aplicar en la enseñanza el reconocimiento de ciertas diferencias entre el español (lengua

materna) y el inglés (lengua meta) para anticipar y remediar errores de interferencia y de
adquisición universales (simplificación y sobre generalización).

4. Familiarizar a los estudiantes con los conceptos y terminología básicos de la gramática, el

vocabulario y la fonología para hablar en inglés.

También se pretende que los estudiantes, además de apreciar las semejanzas y diferencias más
marcadas entre el español y el inglés, aprendan a aplicar estos conocimientos en su futuro trabajo
como maestros de inglés, y tengan en cuenta la naturaleza y el papel de los errores, tanto los de
interferencia como los de adquisición, en el aprendizaje de una lengua extranjera.

En las orientaciones didácticas se dan algunos ejemplos de actividades a realizar, para observar las
semejanzas y diferencias entre el español y el inglés, y desarrollar la habilidad de aplicar el análisis
comparativo y el análisis de errores en la enseñanza del inglés. En el Anexo se dan ejemplos de
materiales y actividades específicos.

CONTENIDO Y ORGANIZACIÓN DEL CURSO

Se cubrirán los siguientes aspectos:

1. Introducción al análisis comparativo y al análisis de errores y al conocimiento del papel que

juegan en la enseñanza del inglés.

2. Análisis comparativo de la gramática del español y del inglés, destacando las semejanzas y
diferencias más significativas para la enseñanza y el aprendizaje del inglés.

3. Análisis comparativo del vocabulario del español y del inglés, destacando las semejanzas y
diferencias más significativas para la enseñanza y el aprendizaje del inglés.

4. Análisis comparativo de la fonología del español y del inglés, destacando las semejanzas y
diferencias más significativas para la enseñanza y el aprendizaje del inglés.

5. La anticipación de dificultades y errores en la planeación de clases.

ORIENTACIONES DIDÁCTICAS

Como actividad introductoria al curso se trabajarán, con algunos extractos de diferentes materiales,
el análisis comparativo y el análisis de errores, y su relación con la enseñanza y el aprendizaje de
las lenguas. A su vez, pueden explorar algunas de las semejanzas y diferencias entre el español y el
inglés (véase Anexo).

En el análisis comparativo de la gramática del español y del inglés conviene subrayar, antes que
nada, que estas dos lenguas se parecen mucho más que, por ejemplo, el español al náhuatl, al
árabe o al ruso. Cuando los estudiantes de inglés se confunden con las estructuras verbales
complejas, muchas veces producen frases que no pueden traducirse literalmente al español, por
ejemplo: He must had being working, que se traduce como Él debe habido estando trabajando, una
oración incorrecta; la oración correcta sería: He must have been working, que en español diría: Él
2
debe haber estado trabajando. Los estudiantes evitarían muchos errores si construyeran las frases
verbales en inglés dentro de lo permisible en español. Sin embargo, hay que recordar que sólo al
escribir o al hacer un ejercicio formal se construyen las frases y oraciones conscientemente; en la
comunicación oral deben salir fluidamente del subconsciente.

Se pueden destacar algunas diferencias generalizables, como:

• La sintaxis. La secuencia de palabras en una frase o un enunciado suele ser menos flexible y
más significativa en el inglés que en el español. Por ejemplo: Has Anne arrived yet? vs. ¿Ya
llegó Anne/Anne ya llegó / Anne llegó ya?
• Los verbos auxiliares. El inglés los emplea más en las estructuras verbales (el español usa
más inflexiones, por ejemplo: Will John come? vs. ¿Vendrá John?), en las respuestas cortas
(como por ejemplo: A: Would you like one. B: Yes, I would), entre otros.
• El género gramatical, concordancia e inflexiones. El español tiene sistemas más complejos,
por ejemplo: Trabajo en varias escuelas y en dos institutos tecnológicos vs. I work in
several schools and in two technological institutes.

Se pueden destacar más las diferencias que influyen en el aprendizaje del inglés de parte de los
hispanohablantes, por ejemplo, el hecho de que el español tenga dos verbos que cumplen funciones
del verbo be le causa problemas a los anglohablantes que aprenden español y no a los
hispanohablantes que aprenden inglés. En cambio, el que el inglés tenga dos verbos que cumplen
funciones para el verbo hacer, sí le causa problemas a los hispanohablantes que aprenden inglés.
Algunas de las diferencias que se pueden anotar son:
• La secuencia verbo-sujeto que se utiliza en afirmativo en español (Llegó mi tío) vs. sujeto-
verbo en inglés (My uncle arrived no Arrived my uncle).
• La omisión de pronombres como sujetos en español (Es muy interesante), que no es
permisible en inglés (It’s interesting, no Is interesting).
• El uso de a antes de un complemento personal en español (Vi a Juan) que no tiene
equivalente en inglés (I saw Juan, no I saw to Juan).
• La secuencia sustantivo-adjetivo del español y su concordancia de número (dos libros azules)
vs. adjetivo-sustantivo en inglés y sin concordancia (two blue books, no two books blues).
• La secuencia adverbio-complemento del español (Hablas bien el inglés) vs. complemento-
adverbio en inglés (You speak English well, no You speak well English).
• El “doble negativo” del español (No hice nada) vs. las opciones complejas del inglés,
nothing/not-anything (I did nothing/didn’t do anything, no I didn’t do nothing).
• Expresiones con tener en español (tener hambre/frío/X años) vs. be en inglés (be hungry /
cold / X years old, no have hungry / cold / X years).
• El verbo gustar (Me gusta Nicole Kidman) vs. like (I like Nicole Kidman, no Nicole Kidman
likes me –a menos que sea cierto).
• El futuro o el presente y el subjuntivo para el futuro en español (Lo haré / hago cuando
lleguemos) vs. el futuro con will y el presente en inglés (I’ll do it when we arrive, no (I will
do it when we’ll arrive).
• El uso del presente y desde hace en español (Vivo aquí desde hace dos años) vs. el presente
perfecto y for en inglés (I’ve lived here for two years y no I live here since two years ago).
• El uso de ser o estar con el participio pasado en español (Soy aburrido vs. Estoy aburrido)
vs. be con el gerundio o el participio pasado en inglés (I’m boring / interesting / etcétera vs.
I’m bored / interested / etcétera).
• El uso del artículo definido con sustantivos genéricos plurales en español (Los pingüinos no
vuelan) vs. la omisión del artículo en inglés (Penguins don’t fly, no The penguins don’t fly)
• El uso de lo + adjetivo en español (Lo importante es...) vs. the + adjetivo + thing en inglés
(The important thing is..., no The important is...).
• La concordancia de los adjetivos / pronombres posesivos con la posesión además del
poseedor en español (¿Juan fue su esposo / esposo suyo?) vs. concordancia únicamente con
el poseedor en inglés (Was Juan her husband? y no Was Juan his husband?). Esta parte se
complica más con el uso en español de su / sus, que corresponden en inglés a your / his /
her / their.
• El uso del infinitivo como sustantivo en español (Fumar es peligroso) vs. el gerundio en
inglés (Smoking is dangerous, no To smoke is dangerous).

3
 • El uso de más como único adverbio de comparación en español (más importante / grande)
vs. More / -er / most / -est en inglés (more important / bigger / most important / biggest,
no more bigger / most big / etcétera).

Estos últimos errores, más que interferencia de la lengua materna, pueden considerarse productos
de un sistema complejo que causa problemas a los niños que aprenden inglés como lengua materna
y de igual forma a quienes lo aprenden como lengua extranjera. El uso de la partícula más en
español representa, hasta cierto punto, una semejanza entre las dos lenguas (más-more) y no una
diferencia; lo que puede provocar problemas es que el sistema en inglés es más complejo que en
español. Otros errores típicos que resultan de la complejidad del sistema inglés más que de la
interferencia del español son:
• El sistema do-does-did (errores como: Does he lives here? She didn’t went).
• El gran número de verbos irregulares (errores como: She teached me English).
• Las distintas complementaciones de los verbos (errores como: She enjoys to swim, Let me
to see y I hope see you soon).
• La secuencia de palabras con los “verbos-frase” (errores como: I picked up it).
• El posesivo con –’s/–s’ (errores como: I’m friend’s John).
• Las formas anything/body/where, además de nothing/body/where (errores como: Anybody
understands me, en vez de Nobody understands me).
• Los plurales irregulares de los sustantivos (errores como: childs / childrens).

De hecho, el error clásico de omitir la “s” en la tercera persona singular del presente simple, se
explica mejor como error universal de simplificación que como error de interferencia y es típico en
los niños británicos y norteamericanos. Además, si las formas verbales del español interfirieran con
el inglés, se agregarían inflexiones en las otras personas en lugar de omitir la inflexión en la única
persona que sí la tiene: I live, you lives, he live, we livimos, they liven.

Después de esta enumeración de semejanzas y diferencias entre el español y el inglés, y las áreas
de errores universales en el aprendizaje de la lengua inglesa, se puede apreciar el reto didáctico,
pero ¿cómo manejar todo esto en forma dinámica e involucrar al estudiante, en lugar de
simplemente presentar el material como listas a memorizar? Recomendamos que se presenten las
distintas áreas como tareas de descubrimiento, por ejemplo:

1. Read the two sentences below, I. produced by a Mexican, and II. by an American. Then do the
tasks and answer the questions.

I. He must had being working all the night.

II. He must have been working all night, poor guy.

a) Translate each sentence into Spanish, word for word, e.g.
i) El debe hab...
ii) El debe....
b) Which sentence translates better into Spanish, word for word?
c) Is this complex English verb phrase correctly constructed in a similar way to Spanish or
very differently?
d) What might be a good strategy for a Spanish speaking learner of English to deal with
such verb phrases?

2. Read the two short texts and answer the questions.

La película era tan aburrida que me quedé dormido. Ana también estaba aburrida pero no se
durmió.

The movie was so boring I fell asleep. Ana was also bored but she didn’t go to sleep.
a) Do the two texts convey essentially the same ideas?
b) What form in English corresponds to “era aburrida” in Spanish?
c) What form in English corresponds to “estaba aburrida” in Spanish?
d) What is the difference in meaning between the expressions (Spanish and English)
referred to in b) and in c) above?
e) How is that difference in meaning conveyed 1) in Spanish, and 2) in English?
4
 f) If a Mexican goes into a clothes store and says “Good morning. I’m interesting in a
swimsuit”, what do you think that person really wanted to say?

Un enfoque didáctico parecido, tareas de descubrimiento, puede emplearse también para las áreas
léxica y fonológica.

Los cognados y los falsos cognados tal vez sean algo relevantes en el área léxica, o cuando menos
más fáciles de tratar, aunque en realidad sea más complejo de lo que parece. Por un lado, el hecho
de que el español se deriva principalmente del latín y una parte importante de la lexis del inglés
también se deriva del latín y del francés puede constituir una ayuda en la comprensión, sobre todo
en la de lectura (en la producción puede crear un tono demasiado formal y rebuscado). El peligro,
por supuesto, radica en la interpretación de todas las palabras parecidas en la forma y el significado,
como: actual-actual (en lugar de present / current), lecture-lectura (en vez de conferencia), library-
librería (en lugar de biblioteca), parent-pariente (en vez de padre / madre), etcétera. Sin embargo,
no es suficiente memorizar los falsos cognados más comunes ya que la relación inglés-español
muchas veces no es sencilla y muchas palabras son cognados parciales, según el uso o contexto.
Por ejemplo, current sí corresponde a corriente de aire / electricidad / etcétera, pero the current
director no significa el director corriente, sino el director actual. A veces las redes de
correspondencia son complejas:
• race = raza ‡ race (de automóviles: carrera).
• career = carrera (en el trabajo) ‡ university course.

Aparte de los cognados / falsos cognados / cognados parciales, hay ciertos subsistemas o campos
léxicos que se estructuran de manera distinta en las dos lenguas. Los más notables son los grupos
familiares en plural de género mixto masculino y femenino: parents / mother and father, siblings
(formal) /brothers and sisters, children / sons and daughters, uncles and aunts, etcétera.

Entre las tareas de descubrimiento se puede incluir la corrección de textos supuestamente

producidos por estudiantes mexicanos:
A: What about your family?
B: Well, I have two brothers, Alicia and David.
A: Do they assist to school?
B: Yes. They were in primary last year, but their actual school is Secondary 11.

También se pueden diseñar tareas que requieren del uso de un buen diccionario inglés-inglés o
inglés-español-inglés.

El tratamiento de la fonología comparativa depende mucho del tiempo disponible después de ver la
gramática y el lexis. Primero, conviene subrayar que los problemas no son exclusivamente fonéticos
(la diferencia entre los sonidos supuestamente individuales –aunque raras veces se producen en
aislamiento– y que cuando se acompañan de otros sonidos cambian su carácter). Puede ser
conveniente empezar con el ritmo, énfasis, contracciones, formas débiles y entonación de
enunciados completos en el inglés, por ejemplo: D’y-think-it’s-gonna-rain? (Do you think it is going
to rain?), comparado con la pronunciación más silábica y pareja del español. Es posible notar aquí
que hay ciertos estilos de hablar en español que tienden hacia el estilo normal en inglés: Q’húbole.
¿Pa’qué me hablas? Yo no po’hacer nada. Por supuesto, la concienciación de la pronunciación
normal de enunciados en inglés tiene más importancia en la comprensión que en la producción; uno
puede hablar despacio, “silábicamente”, y ser inteligible, pero también hay que entender a los
interlocutores.

Entre los problemas fonéticos más relevantes para un estudiante mexicano del inglés son:
• La pronunciación ortográfica: pay pronunciado /paI/ en lugar de /peI/. Esto realmente es
una interferencia de la distinta relación entre la ortografía y la pronunciación en las dos
lenguas, porque el español viene siendo más “fonético”.
• La introducción de /e-/ ante palabras que empiezan con /sp-/ /st-/ /sk-/ /sm-, por ejemplo:
/espi:k/ en lugar de /spi:k/ (speak), /esku:l/ en lugar de /sku:l/ (school), entre otros.
• El debilitamiento o la pérdida de consonantes finales, sobre todo /ba/ en lugar de /bæd/
(bad), /dox/ o /do/ en lugar de /dog/ (dog).
• Dificultad para distinguir entre fonemas distintos en inglés cuando sólo hay uno en español:
/I/ vs. /I/ - sheep vs. ship.
5
 /æ/ vs. /a:/ vs. / / - cam vs. calm vs. come.
/ / vs. /o/ vs. /o:/ - nut vs. knot vs. nought.
/t/ vs. /T/ - tree vs. three.
/b/ vs. /v/ - best vs. vest.
/ƒ/ vs. /tƒ/ - wash vs. watch.

La distorsión de la pronunciación del inglés por la interferencia de los hábitos de pronunciación en

español puede crear un acento tan fuerte que es difícil de entender, salvo entre los que hablan igual
u otros estudiantes con el mismo acento marcado.

Las principales actividades para manejar esta parte del curso probablemente serán:

1. Discriminar entre palabras parecidas.

Teacher: Listen, a) bad, bad; b) bath, bath. Again: a) bad; b) bath. Now, decide which it is, a)
d, or b) th. Mad (math, hid, mud, broth, etcetera).
2. Producir palabras parecidas.
Teacher: OK, look at the words and sentences on the board: a) vest, b) best, c) curve, d) curb,
e) I have trouble with my vowels, f) I have trouble with my bowels. Now say the one I indicate.
Clause b) Sandra.

El último componente del curso, la anticipación de dificultades y errores en la planeación de clases,

es probablemente el más importante, cuando menos en relación con la formación de los estudiantes
como maestros de inglés. La actividad principal será la planeación del contenido lingüístico de una
clase –la función, la gramática, etcétera– y los exponentes o modelos específicos. Por ejemplo:

1. You are planning a class about “The Family”. What grammatical structures are you going to use?
[Verbs: present tense of be, have and basic activities, e.g. work, study. Nouns: mother, father,
parents, brother, sister. Possessives: my, your, his, her, their.]
2. What actual types of statement, question and answer will you use?
[Do you have any brothers or sisters? Yes, I have.../No, I don’t. What are their names/is his/her
name? Do they work or study? Where are your parents from?, etcetera.]
3. What grammar problems do you anticipate?
[Confusion or omission of do/does: You have a brother? Do he works? Omission of 3rd person –s,
or use of it in the interrogative: He study. Where does he studies? Confusion of your/his/her:
What is your name? (about a brother or sister, i.e. his/her name), etcetera.]
4. What vocabulary problems do you anticipate?
Use of brothers for brothers and sisters. Use of fathers for parents.]
5. How would you pronounce the questions Do you have any brothers or sisters? and What is his
name?
[D’you have any brothers or sisters? What’s (h)is name?]

Finalmente, recordar que también se debe “familiarizar a los estudiantes con los conceptos y
terminología básicos de la gramática, el vocabulario y la fonología para hablar en inglés”. Mucho de
esto se expondrá automáticamente al tratar cada tema: subject, cognate, phoneme. Sin embargo,
hay conceptos y terminología que no corresponden de una lengua a la otra, por ejemplo, el inglés
tiene tiempos verbales distintos al español, y un complement es distinto a un object (en She won
the race. She was the champion, the race es el object de la primera oración y the champion es el
complement de la segunda).

Evaluación

Como siempre, la evaluación debe relacionarse estrechamente con las actividades del curso. Por lo
tanto se recomienda emplear los mismos tipos de tarea, sobre todo los indicados para trabajar en la
gramática, el lexis y la planeación de clases. Se podría asignar calificaciones a las mismas tareas del
curso a manera de evaluación continua, o si se requiere un prueba final formal, ésta podría consistir
en una serie de breves tareas del mismo tipo.

6
 Bibliografía general

• James Carl (1980), Contrastive Analysis, Harlow Essex, Longman Group Ltd
• Norrish John (1983), Language learners and their errors, Hong Kong, Macmillan Press
• Wallace Robinett Betty & Schachter Jacquelyn (1983), Second Language Learning, U.S.A.,
The University of Michigan Press
• Inter American University of Puerto Rico (1973), Readings in Spanish-English Contrastive
Linguistics, San Juan, Puerto Rico, Inter American University Press,
• Roca Iggy and Johnson Wyn (1999), A course in Phonology, Great Britain, Blackwell
Publishers Inc.
• James Carl (1980), Contrastive Analysis, Harlow Essex, Longman Group Ltd
• Wallace Robinett Betty & Schachter Jacquelyn (1983), Second Language Learning, U.S.A.,
The University of Michigan Press
• Páginas Web
• http://www.bilingualquestions.org/qa207.htm
• http://www.nichd.nih.gov/RFA/hd-99-012/hd-99-012.htm

ANEXO

ACTIVITY 1

Task 1. Discuss this question in groups: Are most errors made by learners of English as a foreign
language a consequence of interference (or transfer) from their first language?

Task 2. Read the extract below. In groups decide whether the passage suggests it is now generally
believed that:
a) Most errors are a consequence of first language interference.
b) Most errors are not a consequence of first language interference.

Until the late 1960s, most people regarded second language learners’ speech as an incorrect version
of the target language. Their errors were believed to be the result mainly of transfer from their first
language. Contrastive analysis was the basis for identifying differences between the first and second
language and for predicting areas of potential error. So, for example, one might predict that a
speaker of French would be likely to express the idea of being cold as “I have cold” in English
because this would be a direct translation of the way this meaning is expressed in French (J’ai froid).
And, indeed, some errors of this type do occur in learners’ language.
Lightbown & Spada, 2000: 72.

Task 3. Translate the first three sentences (Until the late 1960s... areas of potential error) into
Spanish.

Task 4. Without referring to the original English, translate your Spanish version back into English.

Task 5. Compare the original English version and your new version. Are any differences a result of
the influence of Spanish in your English, or of something else?

ACTIVITY 2

Task 1. In groups, discuss the kind of typical errors that Mexican children make while learning
Spanish (e.g. Lo poní en la mesa. No cabo.). Do you think foreign learners of Spanish make the
same kinds of errors?

Task 2. Read the extract below. In groups, decide whether the passage suggests that foreign
learners’ errors are:
a) Mostly a consequence of struggling to discover the structure of the language, just like little
children.
b) Mostly a consequence of first language interference.
c) Partly first language interference and partly struggling to discover the structure of the language.
7
...however, not all errors made by second language learners can be explained in terms of first
language transfer alone. A number of studies show that many errors can be explained better in
terms of learners’ attempts to discover the structure of the language being learned rather than an
attempt to transfer patterns of their first language. Furthermore, some of the errors are remarkably
similar to the kinds of errors made by young first language learners. An example in English would be
the use of a regular -ed past tense ending on an irregular verb (as in... ‘I buyed a bus ticket’).

Lightbown & Spada, 2000: 72.

Task 3. Translate the sentence “A number of studies show... their first language” into Spanish word
for word: “Un número de estudios...”.

Task 4. In pairs, discuss what this bad, word-for-word translation tells you about English compared
with Spanish.

BLOQUE I
INTRODUCCIÓN A LA GRAMÁTICA COMPARADA

PROPÓSITO:

Familiarizar al normalista con los términos y conceptos básicos de la gramática comparada en

general contribuyendo simultáneamente al desarrollo del dominio integral del inglés.

1. La Gramática comparada como medio de predicción y detección de errores en el aprendizaje de

una lengua extranjera.
• Fonología comparativa

Introducir a los alumnos a un conocimiento general de la fonología y las diferencias de los sonidos
entre el inglés y el español.
• Gramática
• Lexis

2. Analizar diferentes teorías y conceptos relacionados con los errores que se cometen al aprender
una lengua
• Conductismo
• Teoría cognoscitiva
• Teoría de interferencia
• Interlenguaje

3. Tipos de errores
• Errores de interferencia
• Errores de adquisición
• Sobregeneralización
• Aplicación incompleta de reglas

BIBLIOGRAFÍA BÁSICA:

• James Carl (1980), Contrastive Analysis, Harlow Essex, Longman Group Ltd,1-97
• Norrish John (1983), Language learners and their errors, Hong Kong, Macmillan Press, pp. 28-
42, 126-130
• Wallace Robinett Betty & Schachter Jacquelyn (1983), Second Language Learning, U.S.A., The
University of Michigan Press, pp. 6-14, 20-31, 73-85, 109-117, 173-190.

8
ACTIVIDADES SUGERIDAS:

1. Análisis práctico de gramática comparada

Introducir a los alumnos al contenido de esta materia a través de la práctica, no la teoría:
a) Leer una carta o cualquier texto escrito en inglés por una persona cuyo primer idioma sea el
español.
b) Localizar los errores gramaticales, ortográficos y léxicos.
c) Tratar de encontrar la razón por la cual el estudiante del idioma inglés cometió tales errores.
d) Concluir la actividad con una explicación breve acerca de los contenidos de esta materia.

2. Análisis teórico

Para dinamizar el estudio de la teoría de esta asignatura, se sugiere presentar una cita de alguno de
los autores incluidos en la bibliografía y fomentar la participación de los estudiantes con el objetivo
de acordar o estar en desacuerdo con la misma y finalmente llegar a una conclusión grupal.

3. Debate

Dividir el grupo en equipos y asignarle una teoría de los errores que se cometen al aprender una
lengua y llevar a cabo la presentación en forma de debate, en el cual cada equipo defienda su teoría
como si fuera la más certera y completa.
4. Grabaciones
Analizar una grabación de estudiantes de inglés cuya lengua nativa sea el español y determinar el
tipo de errores cometidos a través de la gramática comparada.

BLOQUE II
ANÁLISIS COMPARATIVO DEL ESPAÑOL Y DEL INGLÉS

PROPÓSITO:

Concienciar al estudiante de las semejanzas y diferencias más significativas entre el español y el

inglés.

Gramática del español y del inglés

• Sintaxis
• Doble negación
• Uso del artículo definido

1. Vocabulario del español y del inglés

• Cognados
• Falsos cognados

2. Fonología del español y del inglés

• Diferencias fonológicas entre el español y el inglés
• Interferencia ortográfica en la pronunciación
• Problemas fonéticos que enfrenta el estudiante

BIBLIOGRAFÍA BÁSICA:

• Inter American University of Puerto Rico (1973), Readings in Spanish-English Contrastive

Linguistics, San Juan, Puerto Rico, Inter American University Press, pp. 29-30, 44-69, 123-
135, 151-155, 176-196.

9
 • Roca Iggy and Johnson Wyn (1999), A course in Phonology, Great Britain, Blackwell
Publishers Inc., pp. second page before the table of contents and first page after the index of
subjects.

ACTIVIDADES SUGERIDAS:

1. Realizar una actividad siguiendo los siguientes pasos:

a) Solicitar a los alumnos una investigación acerca de los cognados y los falsos cognados.
b) Realizar presentaciones individuales o en equipo.
c) Enlistar los cognados y falsos cognados en el pizarrón.
d) Definir el significado de los mismos o, en su caso, tratar de adivinarlo.

2. Identificar las diferencias entre los sonidos producidos en inglés y los producidos en español con
ejemplos. Se puede aprovechar este tema para introducir el IPA o Alfabeto Fonético
Internacional.

3. Realizar ejercicios de gramática en los que los estudiantes perciban la diferencia entre el español
y el inglés.

4. Estudiar las diferencias en los proverbios, dichos, o expresiones que sean equivalentes en
español y en inglés. Identificar aquellos en los que se conserve el significado pero la forma sea
diferente. Por ejemplo: “tan fresco como una lechuga” equivale a “as fresh as a cucumber”, lo
cual presenta una diferencia léxica.

5. Examinar la traducción de un traductor famoso. Contrastar la traducción con el original y tratar

de identificar las diferencias más significativas en el léxico y la gramática principalmente.

6. Analizar las diferencias entre canciones que tienen versión en inglés y en español. Se sugiere
llevar la música al aula e invitar a los estudiantes a cantar.

BLOQUE III
USO PEDAGÓGICO DE LA GRAMÁTICA COMPARADA

PROPÓSITO:

Aplicar estos conocimientos como bases para el futuro trabajo del normalista como maestro de
inglés.
1. La importancia de los errores del estudiante.
2. Anticipar y remediar errores de interferencia y de adquisición.
3. Discusión de posibles errores que los estudiantes cometen al aprender el idioma inglés.
4. Anticipación de dificultades y errores en la planeación de clases.
5. Predicción de errores y planeación de actividades para eliminar los mismos.

BIBLIOGRAFÍA BÁSICA:

• James Carl (1980), Contrastive Analysis, Harlow Essex, Longman Group Ltd,141-165
• Wallace Robinett Betty & Schachter Jacquelyn (1983), Second Language Learning, U.S.A.,
The University of Michigan Press, pp.148-157, 158-161
• Páginas Web
• http://www.bilingualquestions.org/qa207.htm
• http://www.nichd.nih.gov/RFA/hd-99-012/hd-99-012.htm

10
ACTIVIDADES SUGERIDAS:

1. Estudiar casos de otros maestros que realizan un análisis comparativo del aprendizaje de la
segunda lengua de sus estudiantes. (Se pueden encontrar muchos de estos artículos en
Internet).

2. Proyecto final
• Desarrollar un proyecto final en equipo en el que el normalista ponga en práctica los
conocimientos adquiridos durante este curso.
• El proyecto consiste en analizar la escritura y el habla de 5 estudiantes (como mínimo) de la
lengua inglesa cuya lengua materna sea el español.
• El propósito es comparar el español y el inglés de los estudiantes en los tres niveles en los
que se enfocó el curso: gramático, léxico, y fonológico.
• Llevar a cabo una presentación acerca de los resultados de este proyecto ante todo el grupo.

3. Escribir una lista de ejemplos de errores que pueden cometer los estudiantes al aprender inglés,
tratando de predecirlos siguiendo la teoría del análisis comparativo o gramática comparada del
inglés y el español.

11
12
MATERIAL

DE

APOYO

13
14
 LANGUAGE LEARNERS AND THEIR ERRORS
John Norrish
actually to contrast the structures. This is no
OTHER CURRENT THEORIES OF THE easy task, since it is extremely difficult to
CAUSES OF ERROR devise a consistent system or basis for
contrast. For example, look at this English
CONTRASTIVE ANALYSIS sentence with the Chinese translation
underneath. The literal translation of the
In 1967 Politzer enthused: ‘Perhaps the least Chinese is given in italics.
questioned and least questionable application
of linguistics is the contribution of contrastive Where is the railway station?
analysis.’ (For a recent review of the subject, Huo che jan tzai na li?
see James 1980.) It was believed that by Fire cart stop at which side
comparing two linguistic systems, that of the
mother tongue and that of the target From this sentence it will be noticed that there
language, it was possible to predict areas of are a number of structural differences
difficulty, and thus errors. This theory is between the two languages. The main one is
related to the notion of ‘interference’ that there is no verb equivalent in the Chinese
mentioned above. It was found that, contrary sentence for ‘is’. Then the two ‘words’ for
to expectations, not all of the areas of ‘where’ appear at the end of the sentenced,
difference between two language systems not the beginning. And so on and so on… it
actually resulted in errors. Areas where no can be seen from this very limited example
difficulty had been predicted did produced that to make any kind of contrastive study of
errors. The main problem with contrastive two languages is an enormous task, and one,
analysis seems to be that while parts of two moreover, which has very seldom been
language systems may or may not differ, this carried out. One example of this type of work
does not tell us much about how a learner will is Stockwell, Bowen and Martin’s well-known
go about the learning task. Nor does it work on Spanish and English (Stockwell,
account for the well-attested fact that the Bowen and Martin 1965).
same errors are made by first language
speakers from very differing language However, for the classroom teacher,
backgrounds. For example, the learners who contrastive analysis is not entirely without
made these errors had different backgrounds, use. Many teachers of English will have a good
one spoke an African language as his mother enough knowledge of the two languages (the
tongue, the other a European language: students’ mother tongue and the target
language) to become familiar with certain
When I reached home, I kiss him. characteristic errors that students make, and,
if there are any, parallel forms in the mother
When the evening came, we go to the tongue. A suitable treatment of errors arising
pictures. from translation from the learners L1 might,
at an intermediate or advanced level, involve
Teachers who have experience in different pointing out that while in the mother tongue it
countries, or of teaching students with is possible to say something in this way,
different first languages in the same class, will nevertheless in the target language it is not.
have noticed that similar errors tend to recur, The danger of this approach is that it can lead
virtually independently of the students’ to an analytical teaching style, which as its
mother tongue. These errors involving the prime aim seeks to eliminate certain errors
verb ‘to know’ were made by different rather than to teach communication through
students in a multilingual class: the target language.

Does he know to find the way? Contrastive analysis can, especially in the field
Does she know to play hockey? of pronunciation, indicate with fair probability
certain areas of difficulty. For example, in
The other problem, perhaps even greater than Thai, some words begin with the ng sound, as
the failure of contrastive analysis to predict in English sing. In English this sound does not
errors, is the sheer magnitude of the task of occur at the beginnings of words. Therefore
writing any contrastive analysis of two this might prove a problem for English
languages. In theory you would need as near speakers learning Thai. Note that this is only a
a complete structural description of both possibility; some learners might not find this
languages as possible. Then, it is necessary particularly difficult.
15
Contrastive analysis can be regarded then by (1972), as well as other workers in the field,
teachers as one of a number of devices in touches on. J Richards terms it
their study of learners’ errors. It has its ‘overgeneralization’ and H V George
problems, though; the main one being that ‘redundancy reduction’. On the basis of his
anyone using the technique needs to know experience of the language, the learner
both the mother tongue and the target constructs a deviant structure, for example
language and would also need a good A We are visit the zoo.
grounding in grammar. B She must goes.
C Yesterday I walk to the shop and buy…
General order of difficulty
As Richards points out, this type of error can
One interesting idea which is receiving be regarded as a blend of two structures in
attention from researchers in studying the the ‘standard version’ of the language. The
causes of errors is the ‘General Order of error might be made as a result of blending
Difficulty? Theory (see Richards and Sampson structures learnt early in the learning
and work by Raven, both in Richards 1974). sequence. In the three sentences given as an
Researchers have found that it is difficult for example, sentence (a) shows a blending of
native speakers as well as for learners of the continuous and the simple present and in
English as a foreign language to distinguish sentence (b) both the modal verb and the
between the English sounds /ν/ and /ð/, and standard third person singular –s suffix are
/f/ and /θ/. These distinctions are among the used. Sentence (c) is slightly different, in that
last made by English-speaking children the redundancy (the additional information
learning their mother tongue. As far as which any natural language incorporates) is
structural difficulties are concerned, Carol removed: the adverbial marker ‘yesterday’ is,
Chomsky in an investigation of children’s for the learner, sufficient to indicate a time
control of their mother tongue (English, in this reference, and consequently the –ed is
case), between the ages of five and ten years, omitted from the stem of the verbs. In the
indicates that even by the age of ten a first two examples, (a) and (b), the
considerable proportion of children were overgeneralization is that of removing the
unable to understand the (apparently) simple necessity for concord, and overgeneralizing
structure ‘John asked Bill what to do’ the rule which states that in the present
(Chomsky 1969). She suggests that there is, simple there are no suffixes except for the
regardless of the age by which a child has third person singular. In the case of (c) the
learnt a particular structure, a characteristic redundancy involves the –ed form and
order of learning which is almost invariable. ‘yesterday’, both indicating time past. The
This cannot be simply related to the child’s information in the message is, under optimum
need to express a particular concept, or the conditions, not interfered with – but with less
frequency of use of the structure, as evidence favorable conditions, the listener would have
from different languages shows different only one indicator of ‘time past’ and could
characteristic orders. Recent work on learners miss it, thus leading to failure to interpret the
of English as a foreign language has indicated speaker’s or writer’s intention.
that this apparent hierarchy of difficulty may
explain, at least partly, some of the learner’s What actually gives rise to the
errors in English. overgeneralization can be any one or more of
a number of factors. Some possibilities are the
Indeed, experiments have shown some quite manner or order in which the language items
surprising similarities in achievement between are presented by the teacher or the text; and
different groups of language learners, both the actual exercises which the learner is called
first and foreign. Bailey, Madden and Krashen upon to complete. For example, learners may
(1974), examining results from several produce the following incorrect responses:
experiments, show that both adults and
children who are learning English as a second Exercise Response
language perform very similarly to each other,
although the adult level of performance was She goes: (must)
not so high as that of the children. She must goes.

Overgeneralization I walk to the shop: (Yesterday) Yesterday I

walk to the shop.
Another approach to the explanation of The general pedagogic dictum of ‘never teach
learners’ errors is that which H V George together what can be confused’ is often a
16
sound one. ‘He sings’ is often contrasted with It is probably even more difficult to avoid
‘he is singing’ in the lessons and texts, after errors arising from ignorance of rule
what seems like sufficient learning time. Soon restriction than it is to avoid false
after this the learner produces a blend of the conceptualization. This is because such errors
two ‘he is sings’. often involve the construction of false
analogies, a very similar activity to what
Incomplete application of rules children do when experimenting with their
own language. A learner, for example, may
In addition to overgeneralization, most have cause to use the noun ‘discussion’, and
teachers will be familiar with the reverse side recalls that it is linked to another noun or
of the coin: incomplete application of rules. noun phrase with the preposition ‘about’ (e.g.,
Richards suggests two possible causes here. ‘a discussion about nuclear energy’). What
One is the use of questions in the classroom, then is more natural when the verb ‘discuss’
where the learner is encouraged to repeat the occurs, than to use it with the same
question or part of it in the answer, for preposition, leading to ‘We discussed about
example the oil crisis’? Or similarly, ‘Tell him to write
the letter’ or ‘Ask him to write the letter’ may
Teacher Do you read much? lead to ‘Make him to writhe the letter’.
Student Yes, I read much
The suggestion made before –not to practice
Or together things that can be confused- will not
completely solve this problem; as we have
Teacher Ask her where she lives. seen, the learner is a far more active
Student Where you (she) live(s)? participant in the language-learning activity
than we may have imagined. Once again it
The other possible cause is the fact that the seems that errors are virtually inevitable.
learner may discover that he can
communicate perfectly adequately using Errors as a part of language creativity
deviant forms.
Learners who are limited in their opportunities
Material-induced errors of listening to examples of the target
language tend to form hypothetical rules
Two further type of error which may be about the new language on insufficient
induced by teaching materials are (a) the evidence. Learners need to create new
‘false concept’ and (b) ignorance of rule utterances, but with limited experience of the
restrictions. An example of (a), which will be target language, they may make mistakes.
familiar to many teachers of English, is the
use of the present continuous tense in the This notice was seen in a hotel in China: ‘A
wrong situation. It is not uncommon in English doctor is available for emergent visits’. The
teaching materials to see a series of pictures person who wrote the notice was aware of
illustrating a sequence of actions, with the adjectival forms like ‘urgent’ versus the
caption in the present continuous although the nominal ‘urgency’, ‘delinquent’ versus
use of the text in this context is unnatural. A delinquency’. He either knew the word
more appropriate context would be a radio ‘emergency’ or found it in the dictionary,
commentary of a football match or a detective marked as a noun. His limited English did not
reporting over the phone the actions of a reach to ‘noun plus noun’ compounds to
suspect, for example. produce the correct form ‘emergency visits’.
He played safe, in one sense, and formed
The problem with contrived use of language what would seem, on the face of it, an entirely
items is precisely that since they form data regular form which should fit into the pattern:
which the learner will use to form his ’an’ + adjective + ‘visit’. It is, of course, in the
hypotheses, the learner may be misled in his nature of language to be unpredictable in
assumptions. Motivation, naturalness and a certain places, and this happened to be the
sensible context for the language are clearly one. The word ‘emergent’ exists, but it doesn’t
vital, if we accept the view that learners will mean what the writer imagined it did.
use the data presented to them actively, in Nevertheless, the process leading to the error
order to test the use of the language items is clearly a creative one. (Selinker 1972 calls
and form assumptions as to the kind of this an ‘L2 communication strategy’.) It is a
language they are learning. natural activity of the human who interacts
with his environment in the laudable attempt
17
to make sense of it and to form it to his own
ends. Foreign and second language learning

This same creative activity lead a student Interference from another language has
learning German to use the adjective bissing already been mentioned as one of the possible
(which literally means ‘liable to bite’) in the sources of error in language learning. Let us
wrong context. He had seen the adjective now look at the different forms this may take
used in relation to dogs Achtung, bissiger for the learner of English as a foreign
Hund (‘Beware of the dog’) so imagined he language and the learner of English as a
could use it to mean ‘beware of the teacher’ second language. One hears the target
and created the utterance Achtung bissiger language in a classroom, where his only
Lehrer! The teacher could easily have chance to practice the language occurs. The
categorized the utterance as ‘deviant’ on the other lives in a different type of environment
grounds of unusual collocation, but in fact his where he can see English in use around him,
positive response had a great psychological on advertising hoardings for example, and
effect on the student and his subsequent hear it used by his fellow countrymen in
enthusiasm from learning German. This certain situations in his society.
anecdote is intended to illustrate the fact that
one of the main causes of error, though not Foreign language errors
the only one, is precisely that creativity and
adventurousness in students that the alert Let us consider first of all the learner who
and responsive teacher at any level will wish hears his English mainly in the classroom; in
to encourage. other words, the learner of English as a
‘foreign language’. The errors he makes will
It should be made clear at this stage that relate closely to his own formation of an
there are at least two types of creativity in ‘interlanguage’. They may also arise as a
language use. The first type, which is being result of the data he is presented with by the
referred to here, is the ability on the learner textbook, or by the teacher, who may, for
to use the parts of the language that he has example, consistently ‘mispronounce’ a given
learnt in order to say something that he may sound, or constantly make a grammatical
not have heard before; it is precisely the error. We would of course hope not to find
same process which leads the mother tongue actual grammatical errors made by teachers
child to produce the form ‘goed’. The learner or in textbooks, but it is possible that a
is drawing certain conclusions about how the manner or order of presentation may lead to
language behaves, using as evidence what he the formation by the learner of certain ideas
has seen of the target language. It is this type or ‘hypotheses’ which may not accord with the
of behavior that teachers would surely wish to actual facts of the target language. For
encourage among language learners, despite example, some textbooks introduce the
the fact that it may, on occasion, lead to present continuous tense at a very early
‘deviant’ language forms. The problem is that stage. This tends to happen because it seems
when exams are being prepared for, both to be the easiest tense to demonstrate in the
teachers and taught expect a rather more classroom, and one of the most ‘productive’
defensive mode of behavior; it is often tenses in so far as it can be used to teach
thought that it is considerably safer, in an further items of the language. As we
exam which sets out to asses a learner’s mentioned before, the order in which items
control of the language code, to say what can are presented and the exercises on them may
be said safely rather than take risks. cause overgeneralization or confusion. This is
difficult to prove, but given what research
The second type of creativity is that which is shows us about language learning and the
usually covered by the term ‘creative arts’. It usual order f introducing language items in
is quite rare for people to be able to create textbooks, this explanation at least seems
works of literature in a language other than plausible. The point being underlined here is
their own, even if it has been as thoroughly that the error in the foreign language situation
learnt as their mother tongue. Joseph Conrad, really has to be a result of something that
whose first language was Polish, but wrote happens in the classroom, since it is usually
novels in English, and Samuel Beckett who is only here that the learner comers into contact
Irish but writes in French, are clearly not the with the target language.
norm among writers.
Second language errors

18
This, however, is not the case when we differ from that used by the same lecturer
consider the learner of English in an when addressing a policeman or perhaps a
environment where the language is in regular market trader, on the limited occasions when
use outside the classroom. We assume that he uses English in these situations.
the target language inside the classroom, the Furthermore, the English used by one
language of the textbook and the target of the policeman to another would again be
teacher is one of the standard varieties of different. It is an over-simplification to regard
English, British, American, Australian, etc. in the non-standard utterances as ‘wrong’. We
very many countries where English is used as need different criteria. Perhaps something
an official language, and perhaps as a lingua along these lines might suffice: does the
franca (that is, a language of communication addressee understand the speaker’s meaning
for those with no common local language), the quickly and clearly? This is clearly a major
varieties of English heard outside the topic. It raises the whole issue of what we
classroom may be very different from the mean by ‘correctness’ in language.
standard variety which is the target inside. An
example of a country where this is the case is A problem occurs when a listener new to the
Ghana, with over fifty languages for a environment appears, or when a speaker
population of approximately nine million. It is accustomed to one perhaps rather limited
not surprising that, in this situation, English variety of local English moves into a different
outside the classroom in Ghana develops setting, needing the language for different
some special characteristics of both purposes. It is of interest that these problems
pronunciation and structure. These features also occur among first language speakers.
which are not found in any of the standard This was noted in 1972 by Doughty, Pearce
varieties of English could be termed both and Thornton who attributed the silence of
‘errors’ and ‘mistakes’. But, given that the some students in a new environment to their
language functions efficiently as a medium of lack of familiarity with suitable language.
communication, would be entirely justified in
talking about ‘errors’? As with American As far as the classroom teacher is concerned,
English, so Ghanaian English differs in both what causes deviations in the case of second
lexical and structural items from British language learning is precisely the fact that the
English, since there are, naturally, concepts target language is being used outside the
and objects not found in Britain or British classroom for the effects of this to carry over
English. into the school. Indeed, it would seem to be
potentially counter-productive to treat these
For example: commonly used forms as ‘incorrect’ in any
way, since they are everyday currency in the
Guarantee shoes platform-soled shoes, ‘real world’ beyond the school walls. On the
originally sold with a guarantee of quality. other hand, the teacher clearly cannot
encourage the use of this type of local spoken
Chopbar a type of open-air snack bar serving English in, say, the writing of pieces of formal
hot food work. What seems necessary here is an
approach which stresses the relationship of
Firestone tyre pass tyre a very common different types of language to different
advertisement, using ‘pass’ for the situations. It may be asking something new of
comparative and meaning both ‘passes’ and many teachers to admit that in language, as
‘is better than’ other tires. in many other fields of human experience,
there may be no such thing as correct and
The car will pick you at the airport Pick you incorrect answers to every question, but
up. rather, more or less appropriate ones.

In this type of environment there is much In more practical terms, what can the teacher
more difficulty in judging error. However, if a actually do? One way of dealing with this
teacher believes that language is not taught problem might be to gather examples of as
only to pass exams, more variations become many different types of English used in the
acceptable. If we talk of ‘Ghanaian English’, country as possible –perhaps this could be
we are simplifying the issue. There are, as treated as a class or school project. Different
with any regional variety of a language, many media using English for the transmission of
different levels and types of Ghanaian English. information should be included:
the language is used by a lecturer talking to advertisements (both written and, where
his colleagues in a university seminar will possible, spoken on radio and film), English
19
used in newspapers, on labels and packaging
on manufactory products, in new broadcasts The next stage after collecting examples of
and on television, in business and other ‘local varieties of English’ would be some kind
communications. It is important for the class of study of the differences and similarities.
to collect as wide a set of examples as The depth and sophistication of this stage
possible, both in writing and, if possible, on would obviously depend largely both on the
tape. If English is used as a lingua franca, as, age of the students and the formal
say, the language or one of the languages of grammatical knowledge of the teacher. One
the playground, then this too would be useful interesting approach to this type of problem
material. has already been made. Jean Ure’s Bridge
Course (1974) follows a method close to that
An example of a ‘Ghanaian English’ poem, outlined here and also includes some study of
from a book written and published in Ghana. local languages and their use in different
The poems are suggested by ‘mottoes’ on the situations. This overtly more advanced work
buses. (p. 71, No Time to Die, K. G. Kyei and was designed for teacher training level, and
H. Schreckenbach, 1975) was written in Ghana.

I NO BE LIKE YOU Summary

I no be like you,
At all; what! An attempt has been made to describe at
Proper libilibi* man like you. least some of the many possible causes of
Today you go speak so for here, language learners’ errors
Tomorrow you go talk different 1. Carelessness, which is not always the
For there. student’s fault.
Then because of you 2. Interference from the learner’s first
Big trouble come for language is now believed to play a smaller
Between people plenty like that. –but still significant- part in causing
I no be like you, errors.
At all; what! 3. Translation from the first language can
Wife-chaser like you, lead to difficulty where there are no exact
Boast-man like you parallels between the two languages. And
You want everything here what, in any case, is an ‘exact parallel’?
For Ghana foryourself alone. 4. Contrastive analysis can indicate some
You chop* areas of difficulty but is of limited use
You don’t want your brother too because of the problem of analyzing the
For chop some self. languages in sufficient detail and finding a
Selfish man like you framework for comparison.
I no be like you, 5. Even native speakers seem to learn their
At all, what! language according to an order of
Teef man, Burglar man like you, difficulty.
You you hide corner corner for dark time 6. Both native speakers learning their mother
And teef woman shoes tongue and foreign learners tend to make
Teef shirts errors by applying what they already know
Teef Charlie-wote* sandals of the language to a new situation where
People worka hayd buy am the same rules do not apply.
From broken-time* money. 7. The learner may be able to make himself
I no be like you, understood by only applying some of the
At all, what! rules and continue to produce deviant
You you take your big mouth forms.
Like chimpanzee bottom 8. The learner may not interpret the material
For talk foolish foolish talk presented to him in the way intended by
For people back. the teacher or the textbook writer.
I no be like you 9. False analogies may be made due to
At all, what! ignorance of rule restrictions.
10. Language can be unpredictable and the
learner may make errors by using the
Libilibi crooked, devious Charlie-wote language he has already learnt to say or
rubber sandals write something he has not ever heard or
Chop eat broken-time overtime teef steal read.
(From ‘thief’)
20
While the causes of errors made by learners of pronoun refers, for example My
a foreign language may be limited to their father who…
Suitability of language for the
exposure to the target language in the Appropriacy
situation in which it is used.
classroom, learners of a second language that The natural process of getting a
they come into contact with outside the spoken language in everyday life,
Acquire
classroom. involving little or no conscious
learning.
It may appear on reading this chapter that it A school of psychology concerned
Behaviorist primarily with the interpretation
is a wonder that learners ever manage to
of behavior as habits.
produce error-free forms. Luckily for teachers, The systems (grammar, meaning
they do! Code
and sound) of language.
Concerned with the thought
Cognitive
Errors and mistakes in listening and processes.
speaking The company a word keeps ; the
Collocation
usual context of a word.
A procedure used by learners to
There was enough evidence in the chapter Communication
convey their meaning ; it may not
above to make teachers think it is a wonder strategy
be ‘grammatical’.
that language learners ever actually manage A way of learning a language in
Community
to produce linguistically acceptable forms. In language learning
small groups using an informant
this chapter we will look at errors or mistakes and translation.
made in speech. Teachers may then be Development of a mental picture,
Concept formation
image or idea.
surprised that listeners can understand An activity which contrasts the
everyday unscripted speech! Contrastive grammatical, semantic and
analysis phonological systems of two or
In considering errors and mistakes in listening more languages.
and speaking, suggestions will be made for An error which is not immediately
Convert (error)
apparent.
remedial teaching. It will become apparent
To work out the meaning of an
that there is nothing different about remedial Decode utterance by using knowledge of
teaching from any other ‘good teaching the code.
practice’. The important points are that the A language ruled based not on
method adopted for remedial teaching should what an authority believes
allow the learner to approach the problem in a Descriptive (rule) language users should do, but on
observation of what they actually
very different way and not in the way that
do.
caused difficulties first time round; secondly, Deviant structure An ‘incorrect’ grammatical item.
the teacher should avoid making the learner Aimed at discovering the gaps in a
feel that the need for the remedial exercise is Diagnostic
learner’s knowledge
his fault. Error
A systematic deviation from the
accepted code.
Native speaker slips Extinction The dying out of a habit.
An inaccurate idea about a
False concept
(grammatical) rule.
It must be fairly clear to most language Ease and confidence in using
teachers that in natural, unscripted speech Fluency those parts of the code a learner
there are likely to be more of the mistake and has experienced.
lapse types of deviation from an ideal than in Language behavior becoming
writing, unless the writing is the result of work Fossilization fixed at a certain point in
development.
at great speed with little or no time for
An error which affects the
correction. When we speak, even in our Global error
meaning of the whole sentence.
language, we sometimes make mistakes; The theoretical ideas underlying
Habit theory
these are often so minor, and indeed, so behaviorist psychology
common that the listener is not even aware of Hypothesis
An informed guess made with the
them. help of given information.
The language used by the learner
as he progresses from no
Glossary of linguistic terms knowledge at all of the target
Interlanguage
language to a satisfactory
Emphasizing correct grammar, as knowledge. The interlanguage is
Accuracy constantly changing.
sometimes opposed to fluency.
Having at least two meanings, The effects of ‘habits’ formed in
Interference from
thus leaving the listener or reader the speaker’s first language acting
Ambiguity L1
in doubt as to the speaker’s or upon the target language.
writer’s meaning and intention. First language (usually the mother
L1
Antecedent Noun, phrase or clause to which a tongue).
21
 Second language (target irrespective of the number of
L2
language). unstressed syllables between
Use of what the learner knows them.
about his first language to try and Suprasegmental Relating to stress and intonation.
L1 transfer
assist expression in the target A pattern of syllables, each taking
Syllable timing
language. roughly the same amount of time.
A non-systematic deviation from The language which the learner is
Target language
the code due to human limitations learning.
Lapse
such as fatigue, poor memory,
etc.
Processes used by learners (e.g.
Learning strategies
mnemonics) to assist in learning.
A language used for
communication by people with
Lingua franca
different first languages in a given
environment.
An error which only affects the
Local error meaning of the clause in which it
is found.
The meaning conveyed by the
Message
language code.
A non-systematic deviation from
Mistake the language code indicating
incomplete learning.
Devices to assist the memory (for
example, memorizing the
Mnemonics sentence ‘Read out your green
book in verse’ to remember the
colors of the spectrum).
Mother tongue First language.
A learner’s willingness to take
Non-defensive
risks when using the target
language behavior
language.
Overt error An immediately apparent error.
A failure by the learner
Over-
restrictions where appropriate to
generalization
the application of a rule.
Other learners assisting in
Peer checking checking for errors or mistakes,
usually in class.
An examination of both ‘correct’
Performance
and ‘incorrect’ forms used by the
analysis
learners.
More information than is
Redundancy necessary to derive the message
from the code.
A learner’s removal of apparently
Redundancy
unnecessary information in a
reduction
language system.
A movement backward to an
earlier stage of learning; (in
Regression reading) a backward movement of
the eye to cover material already
looked at.
A distinction often made between
situations of language use:
second language is where the
Second language / language in question is used for
foreign language some purpose(s) outside the
classroom; foreign language is
where the target language is not
used outside the classroom.
Relating to the phonemic sound of
Segmental a language, excluding stress and
intonation.
A syllabus planned along
structural lines, but involving
Spiral syllabus
systematic uses of those
structures and constant revision.
Stress timing A pattern of regular stresses,

22
 SECOND LANGUAGE LEARNING
are engaged in language teaching and in
CONTRASTIVE ANALYSIS, ERROR writing language-teaching materials. However,
ANALYSIS, AND RELATED ASPECTS the contrastive analysis hypothesis also raises
many difficulties in practice, so many in fact
Edited by Betty Wallace Robinet and Jacquelyn that one may be tempted to ask whether it is
Schachter really possible to make contrastive analyses.
And even if the answer to that question is a
THE CONTRASTIVE ANALYSIS more or less hesitant affirmative, then one
HYPOTHESIS may well question the value to teachers and
curriculum workers of the results of such
Ronald Wardhaugh analyses.

During the course of their reading, students of Actually the contrastive analyses hypothesis
linguistics encounter a number of very may be stated in two versions, a strong
interesting hypotheses concerning different version and a weak version. In this paper the
aspects of language and language function. claim will be made that the strong version is
One long-lived hypothesis which has attracted quite unrealistic and impracticable, even
considerable attention from time to time –but though it is the one on which those who write
more, it must be added, from psychologists contrastive analyses usually claim to base
and anthropologists than from linguists- is the their work. On the other hand, the weak
Sapir-whorf hypothesis with its claim that the version does have certain possibilities for
structure of a language subtly influences the usefulness. However, even the weak version is
cognitive processes of the speakers of that suspect in some linguistic circles.
language.
It is possible to quote several representative
A much more recent hypothesis, and one statements of what has just been referred to
much more intriguing to linguists today than as the strong version of the contrastive
the Sapir-Whorf hypothesis, is the language- analysis hypothesis. First of all, Lado in the
acquisition device hypotheses proposed by the preface to Linguistics Across Cultures (1957)
generative-transformation-analysts. The writes as follows:
hypothesis is that infants are innately
endowed with the ability to acquire a natural The plan of the book rests on the assumption
language and all they need to set the process that we can predict and describe the patterns
of language acquisition going are natural that will cause difficulty, by comparing
language data. Only by postulating such a systematically the language and culture to be
language-acquisition device can a generative- learned with the native language and culture
transformationalist account for certain of the student (P. vii)
linguistic universal, the ability to learn a first
language with ease, but also, apparently, Lado goes on to cite Fries in support of this
another universal, the inability to learn a proposition. Here is the appropriate quotation
second language after childhood without from Fries’s Teaching and Learning English as
difficulty. Like the Sapir-Whorf hypothesis, the a Foreign Language (1945):
language-acquisition device hypothesis is
extremely intriguing, but it too presents The most efficient materials are those that are
seemingly insurmountable difficulties to based upon a scientific description on the
anyone seeking to device a critical test to language to be learned, carefully compared
prove its truth or falsity. A linguist may accept with a parallel description of the native
the hypotheses because they usefully and language of the learner. (P. 9)
economically explain certain language data
that he wants to explain in terms of a set of More recently in a book edited by Valdman,
axioms he can accept; or he may reject the entitled Trends in Language Teaching (1966),
hypotheses because they reek of mentalism or Banathy, Trager, and Waddle state the strong
subjectivity, or because he prefers a different version of the contrastive analysis hypothesis
set of axioms on which to base his work. as follows:

Still a third hypothesis, and the one which is … the change that has to take place in the
of special interest in this paper, is the language behavior of a foreign language
contrastive analysis hypothesis, a hypothesis student can be equated with the differences
of particular interest to those linguists who between the structure of the student’s native
23
language and culture and that of the target linguists claim they could follow in order to
language and culture. The task of the linguist, achieve definitive results if only there were
the cultural anthropologist, and the sociologist enough time.
is to identify these differences. The task of the
writer of a foreign language teaching program If one looks specifically at how phonological
is to develop materials which will be based on problems have been dealt with in this strong
a statement of these differences; the task of version, he can easily find evidence to support
the foreign language teacher is to be aware of the assertions just made. Many a linguist has
these differences and to be prepared to tech presented contrastive statements of the
them; the task of the students is to learn phonemic systems of two languages without
them. (P. 37) asking whether it is possible to contrast the
phonemic systems of two languages by
the same idea is presented in each of these procedures which attempt to relate and
three statements, the idea that it is possible English p to a French p, because linguists
to contrast the system of one language –the have chosen to symbolize some not well-
grammar, phonology, and lexicon- with the defined similarity between the two languages
system of a second language in order to in the same way, in this case by the letter p,
predict those difficulties which a speaker of or because both p’s are associated with
the second language will have in learning the certain movements of the glottis and lips. The
first language and to construct teaching use of the similarity of the symbols is more
materials to help him learn that language. deceiving than the use of the similarity of
phonemic features. The latter may be justified
An evaluation of this strong version of the to some extent in terms of what will be
contrastive analysis hypothesis suggests that referred to later as the weak version of the
it makes demands of linguistic theory, and, hypothesis, but statements about a language
therefore, of linguists, that they are in no lacking certain phonemes or two languages
position to meet. At the very least this version having the same phonemes are possibly even
demands of linguists that they have available more dangerous than they are naïve. Any
a set of linguistic universals formulated within such statements must ultimately rest on
a comprehensive linguistic theory which deals phonetic evidence, and, if they do, the strong
adequately with the syntax, semantics, and version of the hypothesis is being disregarded
phonology. Furthermore, it requires that they in favor of the weak version. As Weinrich
have a theory of contrastive linguistics into (1953) points out, phonemes are not
which they can plug complete linguistic commensurable across languages; phones,
descriptions of the two languages being individual sounds, are much more
contrasted so as to produce the correct set of manageable, because they do have some
contrasts between the two languages. Ideally, connection with events in the world, in this
linguists should not have to refer at all to case articulatory and acoustic events.
speakers of the two languages under contrast
for either confirmation or disconfirmation of Let us suppose that a linguist contrasts the
the set of contrasts generated by any such allophonic variants described in accounts he
theory of contrastive linguistics. They should finds of the phonological system of two
actually be able to carry out their contrastive languages. Could he then meet the demands
studies quite far removed from speakers of of the strong version? Once again the answer
the two languages, possibly without even must be negative, at least within the present
knowing anything about the two languages in state of linguistic knowledge. Ideally, a
question except what is recorded in the linguist interested in making a contrastive
grammars they are using. Such seems to be analysis would like to be able to take a
the procedure which the strong version of the statement of the allophones of Language A
contrastive analysis hypothesis demands of and say for each one exactly what difficulties
linguists. Stated in this way, the strong a speaker of Language B would have in
version doubtless sounds quite unrealistic, but producing that allophone. However, the
it should be emphasized that most writers of difficulties in the way of doing this are
contrastive analyses try to create the formidable. Are the phonetic statements the
impression that this is the version of the linguist finds sufficiently detailed and of the
hypothesis on which they have based their right kind to be of use: that is, what is the
work –or at least could base their work if adequacy of the phonetic theory and the
absolutely necessary. Here is yet another particular phonetic information at his
instance of a ‘pseudoprocedure’ in linguistics, disposal? Do the descriptions take into
a pseudoprocedure being a procedure which account all the phonological variables that
24
should be taken into account, such as A close reading of most of the contrastive
segmentation, stress, tone, pitch and juncture analyses which are available shows them to
and syllable, morpheme, word and sentence conform to some of the demands made by the
structures: that is, what is the state of the weak version of the theory and not at all to
phonological theory he is using? Does the the demands of the strong version. Even the
linguist have available to him an overall two highly regarded texts on English and
contrastive system within which he can relate Spanish by Stockwell and Bowen, The Sounds
the two languages in terms of mergers, splits, of English and Spanish (1965) and The
zeroes, overdifferentiations, Grammatical Structures of English and
underdifferentiations, re-interpretations, and Spanish (1965), fall into this category. It
so on: that is, what is the state of the appears that Stockwell and Bowen use their
contrastive theory he is employing? In this linguistic knowledge to explain what they
age of linguistic uncertainty the answer to all know from experience to be problems English
of these questions is obvious. speakers have in learning Spanish. The
linguistic theory they use is actually extremely
It seems, therefore, not a little strange, given eclectic and contains insights from generative
all the problems which the strong version of transformational, structural, and paradigmatic
the contrastive analysis hypothesis creates, grammars; nowhere in the texts is there an
that so many linguists claim to use in their obvious attempt to predict errors using an
work. None of them have actually conformed overriding contrastive theory of any power.
to its requirements in such work. However, Even the hierarchy of difficulty which
there have been attempts, some more Stockwell and Bowen establish in the second
successful and some less successful, to use chapter of the Sounds volume is based more
what may be called the weak version of the on their experience and intuition than on an
contrastive analysis hypothesis. In this case, explicit theory for predicting difficulties.
one must offer his own definition of the weak
version, because the literature contains little In recent years there have been two still
or no reference to what linguists have actually different approaches taken to the problems of
done in practice, in contrast to what they have contrastive analysis, both resulting from the
claimed they were doing or could do. current enthusiasm for generative
transformational theory. One of these
The weak version requires of the linguist only approaches dismisses the hypothesis from any
that he uses the best linguistic knowledge consideration at all. This dismissal stems from
available to him in order to account for a strong negative reaction to contrastive
observed difficulties in second language analysis, as, for example, in recent articles by
learning. It does not require what the strong Ritchie (1967) and Wolfe (1967) in Language
version requires, the prediction of those Learning. The second approach attempts to
difficulties and, conversely, of those learning use the generative-transformational model in
points which do not create any difficulties at order to provide some of the necessary
all. The weak version leads to an approach overriding theory to meet either the demands
which makes fewer demands of contrastive of prediction in the strong version or of
theory than does the strong version. It starts explanation in the weak version.
with the evidence provided by linguistic
interference and uses such evidence to The case for dismissal may be stated as
explain the similarities and differences follows: Languages do not differ from each
between systems. There should be no mistake other without limit in unpredictable ways
about the emphasis on systems. In this statements to the contrary notwithstanding.
version systems are important, because there All natural languages have a great deal in
is no regression to any presystemic view of common so that anyone who has learned one
language, nor does the approach result in language already knows a great deal about
merely classifying errors in any way that any other language he must learn. Not only
occurs to the investigator. However, the does he know a great deal about that other
starting point in the contrast is provided by language even before he begins to learn it,
actual evidence from such phenomena as but the deep structures of both languages are
faulty translation, learning difficulties, residual very much alike, so that the actual differences
foreign accents, and so on, and reference is between the two languages are really quite
made to the two systems only in order to superficial. However, to learn the second
explain actually observed interference language –and this is the important point- one
phenomena. must learn the precise way in which that
second language relates the deep structures
25
to its surface structures and their phonetic Many experienced teachers find themselves
representations. Since this way is unique for unable to accept such reasons for rejection of
each language, contrastive analysis can be of the hypothesis. Their experience tells them
little or no help at all in the learning task that a Frenchman is likely to pronounce
because the rules to be internalized are, of English think as sink and a Russian likely to
course, unique. Even though the form and pronounce it as tink, that a Spaniard will
some of the content of the rules to be almost certainly fail to differentiate English bit
acquired might be identical for both from beat, and thet an Englishman learning
languages, the combinations of these for French will tend to pronounce the French word
individual languages are quite idiosyncratic so plume as pleem or ploom. They admit that in
that superficial contrastive statements can in each case they must be prepared to teach the
no way help the learner in his task. whole of the second language to the learner,
but also insist that some parts of the second
Now there is obviously some merit in the language are easier to learn that others, for
above argument. If the underlying vowel no one ever must learn everything about the
system of French is something like the one second language. However, many also admit
Schane outlines in The Sound Pattern of that they do not know in what order learners
English (1968), and the underlying vowel should try to overcome the various difficulties
system of English is something like the one they are observed to have. Should a Spaniard
Chomsky and Halle outline in The Sound learning English learn to differentiate bit from
Pattern of English (1968), and if the speaker beat and bet from bait because of the
of English must somehow internalize the important surface contrasts which he does not
underlying vowel system of French and the make in Spanish? Or should he learn to
fifty or so phonetic realization rules which associate the vowels in such pairs of words as
Schaune gives in order to speak acceptable weep and wept, pale and pallid, type and
French, then one may easily be tempted to typical, tone and tonic, deduce and deduction
reject the whole notion of contrastive analysis, so that he can somehow internalize the
claiming that it has nothing at all to contribute underlying phonological system of English?
to an understanding of the learning task that The mind boggles at this last possibility! But it
is involved. is one which descriptions of Spanish and
English based on generative transformational
Uncertainty is obviously piled upon theory would seem to hold out for teachers.
uncertainty in making contrastive analyses.
Such uncertainties arise from inadequacies in Some recent suggestions for using generative
existing linguistic theories. As an example of transformational theory in contrastive analysis
theoretical inadequacy, one may observe that have actually been attempts to bring powerful
the notion of deep structure itself is extremely theoretical insights to bear within the weaker
uncertain. Chomsky (1968), McCawley version of the hypothesis in interesting work,
(1968), and Fillmore (1968) all mean Ritchie and Carter have used distinctive
somewhat different things by it, but all at feature hierarchies in attempts to explain such
least agree that it has something to do with problems as why a Russian is likely to say tink
meaning. However, for the purposes of and a Frenchman sink for English think, using
contrastive analysis any claim that all the notions of feature hierarchy, rule cycling,
languages are very much the same at the and morpheme and word structure rules, has
level of deep structure seems to be little more considerable possibilities. Certainly this kind of
than a claim that it is possible to talk about work seems more promising than some being
the same things in all languages, which is done by others in an attempt to show gross
surely not a very interesting claim, except similarities between deep structures in an
perhaps in that it seems to contradict the one assortment of languages.
made by Sapir and Whorf. The preceding
statement is not meant to be a criticism of In conclusion, it is fair to say that teachers of
generative transformational theory; it is second and foreign languages are living in
meant to show how acceptance of generative very uncertain times. A decade or so ago
transformational theory; it is meant to show contrastive analysis was still a fairly new and
how acceptance of that theory can fairly easily exciting idea apparently holding great promise
lead one to reject the idea that generative for teaching and curriculum construction.
transformational theory has something to Now, one is not so sure –and not solely as a
contribute to a theory of contrastive analysis, result of the Chomskyan revolution in
given the present state of the art. linguistics. The contrastive analysis hypothesis
has not proved to be workable, at least not in
26
the strong version in which it was originally the existence of ch in both Spanish and
expressed. This version can work only for one English orthographies with approximately the
who is prepared to be quite naïve in linguistic same sound values is a condition for positive
matters. In its weak version, however, it has transfer: the familiar ch of church carries over
proved to be helpful and undoubtedly will to chile, leche, lechuga with positive effect.
continue to be so as linguistic theory Finally the symbol ñ might lead to zero
develops. However, the hypothesis probably transfer –but in fact, since the student is
will have less influence on second language familiar with n but not with ñ, he often ignores
teaching and on course construction in the the tilde and hence encounters negative
next decade than it apparently has had in the transfer. An unarguable instance of zero
last decade. One cannot predict whether that transfer for the reader of English does not
diminishing influence will have a good or bad exist in Spanish orthography; we must look
effect on second language teaching. Today instead to a symbol system like those of
contrastive analysis is only one of many Korean or Chinese to find true instances of
uncertain variables which one must reevaluate zero transfer to him.
in second language teaching. No longer does
it seem to be as important as it once was. The conditions of negative, positive, and zero
Perhaps, like the Sapir-Whorf hypothesis, it transfer by themselves would enable us to set
too is due for a period of quiescence. up a reasonable hierarchy of difficulty. We
could safely assume that instances where
SOUND SYSTEMS IN CONFLICT: A conditions for positive transfer existed would
HIERARCHY OF DIFFICULTY lend themselves to mastery more readily than
instances where conditions for a negative or
Robert P. Stockwell and J. Donald Bowen zero transfer existed. It is probable that we
should have somewhat more difficulty
In attempting to arrive at a reasonable determining whether the instances of negative
hierarchy of difficulty, we must take into transfer were more difficult than those of zero
account information from what psychologists transfer: does the student have more trouble
have developed as learning theory. There are mastering gender concord in Spanish (el
no doubt many aspects of learning theory muchacho mexicano, but la muchacha
from which we might benefit, but one concept mexicana), an instance of zero transfer, or
in particular seems promising: the notion of with por-para, where the phonetic similarity of
transfer –negative transfer, positive transfer, por and English for seemsto set up an
and zero transfer. A student may have some instance of negative transfer? Indeed we
habitual responses which are contrary to the would have no little difficulty deciding exactly
responses required for a new skill which he is which instances involved negative transfer
trying to master (negative) or which are and which ones zero: it is not at all clear, for
similar to the new responses (positive), or example, whether ser-estar is difficult because
which have no relation to them (zero). This of negative transfer from is to es, or because
notion of transfer is applicable throughout the of zero transfer from lack of distinction
structure of the language: the sound system, between such verbs in English to presence of
the grammar, the vocabulary. Let us see it in Spanish, or because of both factors
illustrations of transfer based on the relation together.
of pronunciation to spelling.
It seems that we may get around the
Suppose, for instance, that a student is trying difficulties inherent in the question of types of
to learn to pronounce Spanish by using transfer by focusing our attention on the kinds
Spanish orthography as a guide. He sees the of choices that exist at any given point in the
word Habana, spelled (as in English) with an two languages. We have already seen that the
initial h-. but the h- is ‘silent’ (i.e., represents pronunciation of a language may be
no phonological reality) in Spanish characterized as a set of choices, plus
orthography. The student’s literacy habits obligatory consequences, or, as we might say,
conditioned him to produce the initial sound of optional choices and obligatory choices. We
have, hold, her, him when he sees h-. these can add to these a third set; zero choices –
are the conditions of negative transfer –a those which exist in one language but not al
familiar response to a familiar stimulus is tall in the other. An example is the phoneme
wanted. The effect of the old response is /ž/-the middle consonant of pleasure- which
negative: he pronounces the Spanish word exists in English but not in Spanish. We can
with an h-. on the other hand, to continue now set up the following three-way
with orthographically conditioned transfers,
27
correspondences between English and difficult Spanish would be to teach if this
Spanish. (Op optional, Ob obligatory, Ø zero). set of choices were not held in common.

There are eight possible situations, not 2. English Ob, Spanish Op. examples for this
counting the theoretical ninth possibility of comparison are scarce. If we limit our
zero choice in both languages: coverage of English to a particular dialect,
however, an example can be found. In the
English choice
Spanish dialect that is sometimes called southwest
choice midland (Oklahoma, Arkansas, southern
Op Op
Missouri, southern Kansas, northwest
Ob Op
Texas), the vowels of pin and pen are
Ø OP
Op Ob identical. That is, speakers of this dialect
Ob Ob have no choice between /I/ and /E/ before
Ø Ob /n/. they can of course choose other
Op Ø vowels, like those of pat, pot, bought,
Ob Ø beat, but the only vowel they can choose
in the area of /E/ and /I/ is a vowel which
In this method of comparison of sound is really neither one of these but more or
systems, optional choice refers to the possible less midway between. It is a well-known
selection among phonemes. For example, the joke that they can distinguish between pin
English speaker may begin a word with /p/ or and pen only by specifying a “stickin’ pin”
with /b/. Obligatory choice refers, for one or a “writin’ pin”. For these speakers it is
thing, to the selection of conditioned clear that there is no choice between /E/
allophones. For example, when the English and /I/ in the environment: -n. faced with
speaker has /p/ at the beginning of a word, a Spanish item like lento, the conditions of
the structure of the language requires the negative transfer exist for them: they will
aspirated allophone /p’/ in that environment. regularly produce the only vowel their
Also, obligatory choice refers to limitations in dialect allows in the general phonetic area
distribution of phonemes. For example, before of /E/ or /I/, and it is not very similar to
/m/ at the beginning of q word, English has the correct vowel.
only /s/, never /z/. the term ‘zero choice’,
which is meaningful only when two languages 3. English O Spanish Op. This correspondence
are being compared, refers to the existence of characterizes the classic difficulty the
a certain sound in one language which has no English speaker has with the erre of
counterpart at all in the other. Let us see what Spanish perro, or the jota of Spanish hijo.
sort of examples might exist for each type. In neither instance does the sound exist in
English, although both sounds represent
1. English Op, Spanish Op. both languages optional choices of considerable frequency
allow certain consonants to appear at the in Spanish. From the English speaker’s
beginning of a word before a vowel. There point of view they are new sounds.
are words like me, knee, tea; mí, ní, tí;
and others. We can symbolize this fact in 4. English Op, Spanish Ob. This
a general way: correspondence characterizes one of the
more difficult problems of Spanish
/m phonology for the English learner. Take,
In env. -
English / for example, the pronunciation of items
v
/n/ like dado and dedo in isolation. The d at
/t/ the beginning is pronounced differently
from the d in the middle. The initial d is
That is, initially before a vowel, English much like the initial d of English den, doll,
and Spanish share the possibility of door. (It is not exactly the same, but the
choosing such consonants as /m, n, t/. difference is irrelevant for this purpose).
Although this description is obviously We will write it with the phonetic symbol
incomplete, since the full list of possible /d/. the middle d of dado, dedo, on the
consonants is not specified, the mere fact other hand, is conspicuously different –to
that the two languages share a specifiable the English ear- from the initial d. it
list of prevocalic consonantal possibilities sounds more nearly like the initial th of
is a huge source of positive transfer. One then, there, those. We will write it with the
can barely imagine how much more phonetic symbol /€/. Dado and dedo can
now be written phonetically as /da€o/,
28
/de€o/. for te Spanish speaker the the matter: the occurrence of Spanish /d/
pronunciation of /€/ rather than /d/, in and /€/ can be predicted in writing merely
the middle of these words is obligatory. He one symbol, /d/; given this symbol in an
will not ordinarily even be aware that he environment, it is possible always and
pronounces two quite different sounds for infallibly to predict whether it will be
the d’s of dado and dedo. To use the pronounced /d/ or /€/. The difference
technical terminology introduced earlier, between them is obligatory.
/d/ and /€/ are allophones of a single
phoneme /d/ in Spanish. Among the This correspondence between English
consonants of Spanish, /d/ exists as one optional choices and Spanish obligatory
possible optional choice, which we may be choices is so important in its
symbolized: consequences that another example may
/p/ clarify it still further. Suppose we consider
/t/
the possibilities of nasal consonants (/m/
/k/
Spanish C /b/ In env. –V as in ham, /n/ as in hen, /N/ as in hang) in
/d/ the environment of the following stop
/g/ consonants (/p/ as in up, /t/ as I putt, /k/
as in puck, /b/ as in tub, /d/ as in dud, /g/
There is then a subsidiary rule about /d/ as in dug). The phonetic symbols needed
(illustrated, incompletely, below): for this discussion are all familiar letters f
the alphabet in familiar values, except for
/l/
/d/ in env. /N/. Note that the letters ng are used to
/n/
spell both /N/ and /Ng/ in English: words

/d/ like singer and banging have /N/, whereas
words like finger and younger have /Ng/.
/€/ in env. V Certain articulatory facts about these
consonants must be briefly explained in
That is, if /d/ is preceded by silence (a order to make the point clear. In terms of
break in utterance continuity symbolized the place in the mouth at which the sound
in the formula by ) or an /n/ or /l/, it is is articulated,, the nasal and stop
pronounced as /d/. if it is preceded by a consonants fall into three classes: those
vowel, it is pronounced as /€/. The made at the lips (/m, p, b/); those made
phonetic difference between /d/ and /€/ is by the tip of the tongue at or just behind
conditioned by this rule –a rule which the upper teeth (/n, t, d/); and those
merely describes a set of conditions to made toward the back of the mouth, with
which Spanish speakers habitually, and the tongue touching the back part of the
unconsciously, conform. Because of this roof of the mouth (the velum) (/N, k, g/).
rule, /€/ is for them simply a kind of /d/.
but for the English speaker, the conditions Lips Teeth Velum
are different. For him /d/ and /€/ are in m n N
contrast –that is, they belong to different p d k
b t g
phonemes, /d/ and /€/. The fact of
contrast is proved by pairs such as dine-
It is characteristic of Spanish that in a
thine, dare-there, dough-though. /d/ and
sequence of nasal consonant plus stop
/€/ exist as two possible choices among
consonant the point of articulation of both
the consonants of English:
/p/ consonants is fixed by the stop consonant.
/t/ This can be formulated:
/k/ /p/
/b/ /m/ in env. – /b/
English C /d/ In env. -V
/g/
/v/
/t/
/€/ /n/ in env. – /d/
Spanish N
In English, unlike Spanish, /d/ and /€/ are
in contrast: they are both optional
choices, and their distribution cannot be /k/
/N/ in env. -
predicted. Predictability is at the heart of /g/
29
 That is, a nasal (N) can be only /m/ if the at an early age and has given no thought
following consonant is /p/ or /b/, only /n/ to since. Zero may be viewed as a kind of
if the following consonant is /t/ or /d/, negative obligation: to say that a pattern
only /N/ if the following consonant is /k/ or is zero is about the same as saying that it
/g/. this restriction remains valid is obligatory that the speaker not conform
regardless of word boundaries and to the pattern. We have, as it were, an
spelling: hombre, un beso; endosar, un absolute negative restriction in the one
día; inglés, un gato. It is optional whether instance, an absolute positive restriction in
a nasal be chosen at all; but if one is the other. An example is to be found in
chosen, it is obligatory that its point of the middle consonant of Spanish words
articulation be the same as that of a like haba, leva, avance. Although spelled
following stop consonant. In English, on with b or v, this sound is different from
the other hand, no such restructions exist: anything represented by b or v in English.
/mb/ lumber, /nb/ unbend, /Nb/ kingbird; The phonetic symbol we will use for it is
/md/ lambda, /nd/ under, /Nd/ kingdom; /§/. It is articulated by bringing the lower
/mg/ Baumgardner, /ng/ ingrown, /Ng/ lip up toward the upper lip, as if for b, but
finger. In English, not only is the choice of without touching, so that the air produces
a nasal consonant optional, as in Spanish, a friction noise, as if for v. in Spanish, the
but so is the choice of a particular nasal, difference between /b/ and /§/ is closely
regardless of the following stop consonant, parallel to that between /d/ and /€/. The
which is not true in Spanish. two sounds are allophones of a single
phoneme, predictable from a single
5. English Ob, Spanish Ob. It is here that we symbol in the following way (this
get maximum positive transfer. Any formulation of the rule is illustrative only,
English pattern that is obligatory is not complete):
necessarily one to which the speaker gives /m/
no thought –it is an area where he has no
choice. If the same pattern is obligatory /b/ in env.

also in Spanish, there should be no
problem- indeed, there will not normally Spanish /b/
even be any awareness that there might /§/ in env.
V
have been a problem. These instances are
more frequent than we realize:
comparison between Japanese and
Spanish, on the one hand, and between
English and Spanish, on the other, will That is, /§/ normally occurs after vowels,
reveal that the English speaker is not so /b/ elsewhere. The situation of /b/-/§ is
bad off for Spanish-like habits as we who different from /d/-/€/ in only one
are faced with the student’s errors are significant respect: /§/ does not exist in
prone to think. To take a simple instance: English at all (a zero category), but /€/
given the consonantal sequence /s/ plus does (an optional category). But it is a big
/w/, both languages require that a vowel difference pedagogically. In the instance
be chosen in the next position –swear, of /€/, the English speaker must transfer
suerte. This is not a trivial observation: if a familiar sound and redistribute it with
the consonantal sequence is /p/ followed respect to other sounds; in the instance of
by /r/, English requires a vowel, as in /§/, he must learn a new sound as well as
pray, but Spanish allows /y/ or /w/, fro a new distribution.
example prieto /pryeto/, pruebo /prwébo/.
Thus the fact that the Spanish speaker has 7. English Op, Spanish ¥. This particular
a different range of choice after /pr/ correspondence is a frequent one on going
constitutes a problem for the English from English to Spanish pronunciation.
speaker, even though the sequence /pr/ English has several vowels, for instance,
itself does not. that are entirely lacking in Spanish. The
vowel of American English grass,
6. English Spanish Ob. This correspondence symbolized by /Q/, does not exist in
is the extreme of the scale. In English, a Spanish. Partly because of negative
given habit does not exist at all; in transfer from the spelling a, words like
Spanish, it is obligatory and hence gracias are often pronounced with this
normally outside the speaker’s conscious vowel in early stages of learning. The
control –it is a habit which he internalized problem is merely to reduce the range of
30
 choice that the English speaker is List of
accustomed to exercising. prevocalic
III 7 Op Op 1
consonant
s
8. English Ob, Spanish. An English obligatory Sw-plus
pattern of pronunciation can be difficult to 8 Ob Ob 5
vowel
get rid of. For instance, it is obligatory in
most English dialects that item with t or d Given such a hierarchy, we must examine
between syllables, where the first syllable several other criteria that will enter into the
is stressed (butter, shudder, splatter, grading and sequencing of materials designed
Betty, patty), have an allophone for /t/ (or to eliminate these difficulties.
/d/) that is rather like the Spanish r of
para, pero. It is a voiced tongue-tip flap. The most important of these is functional load
Faced with Spanish words like foto, beta, –that is, the extent to which a given sound is
pita, the English speaker of most dialects used in Spanish to distinguish one word from
will produce the obligatory English flap another, the quantity of distinctive
rather than the fully articulated /t/ of information that it carries. The Spanish ñ
Spanish. Another example also involves belongs in Group I in the hierarchy of difficulty
allophones of English /t/: in items like (in English, optional in Spanish). But its
mountain, button, latent, the English functional load is almost zero. There are about
speaker of most dialects has a variety of a dozen words in which ñ carries the burden
/t/ for which instead of dropping the of contrast with the cluster /ny/ (spelled –ni-
tongue tip as he usually does to release a ): uñón (big toenail) versus union and the
/t/, he maintains the tongue tip in the like. An American can speak Spanish for a
same position for the following /n/. such long time without ever needing this contrast.
an articulation does not exist under any For ñ he can substitute /ny/, modifying a
conditions in any dialect of Spanish. Words cluster he controls from his English habits only
like quitan, meten, which always have a to the extent of being careful not to make the
normally released /t/ and a full vowel, are syllable division between /n/ and /y/ -that is,
subject to this kind of transfer. ha must say /u.nyón/ rather than /un.yón/.
the ñ would, therefore, in spite os itf relatively
Having at least an idea, now, of the eight kind high rank in the hierarchy of difficulty, be
of differences that a comparison can reveal placed very late (indeed, almost last) in a
when it is based on different possibilities of reasonable pedagogical hierarchy.
choice in the two languages, we can attempt
to rearrange the comparisons in an order A less important additional criterion is
which will constitute a hierarchy of difficulty. potential mishearing. Spanish initial /t’/ -the
We must know which kind of differences will variety of /t/ that appears before vowels –
be most difficult to master and which will be provides an example. This sound is very
easiest, in order to grade our teaching difficult. Group I (O in English, obligatory in
materials, arrange them onto an effective Spanish), in our hierarchy above. But failure
sequence, and determine how much drill is to produce it correctly (with the tongue tip
needed on each point. The hierarchy against the back side of the upper teeth,
suggested below is by no means final; further without a puff of air) will rarely cause
experience with it may well result in misunderstanding. However, the American
readjustments in the relative position of one who is listening rather than speaking –
category of difficulty or another. receiving rather than producing- will often
hear a Spanish initial /t’/ as being a /d/. one
Difficulty Comparison Examples good way for him to learn to hear it correctly
from
Magnitud Orde Englis Spanis Typ is for him to produce it correctly. We would
preceding
e r h h e therefore place the /t’/ fairly high in a
discussion
1 O Ob 6 /§/ pedagogically oriented sequence even though
I 2 O Op 3 Erre, jota when evaluated as to its effect on the
3 Op Ob 4 /d/ - /€/ production of Spanish, it will only add
I/e/ before American accent to the student’s
4 Ob Op 2
n
pronunciation –not unintelligibility at any
Flap /t/
II
5 Ob O 8 between
point.
vowels
6 Op O 7 /Q/ The final additional criterion is patter
congruity. The sounds of a language pattern
themselves in groups or sets. In Spanish, /b/,
31
/d/, and /g/ constitute a set. /b/ and /d/ are language (L1) on L2 behavior. i.e., the
high in difficulty, in functional load, and in elimination of interference behavior. it follows
potentiality for mishearing. There is no doubt that we cannot expect to attain maximum
they must appear early in a pedagogical success in the teaching of the practical
sequence. /g/ is also difficult, but it is phonology of an L2 unless we have a clear
considerably lower in functional load and has understanding of what the nature of the
less potential for mishearing. Because it influence of the L1 on L2 behavior might be –
patterns like /b/ and /d/, we feel it would be i.e., unless we have an explanation of the
incongruous to place it out of sequence with interference behavior. this paper suggests
them even though it does not constitute a that certain modes of explanation based on
problem of the same order. conventional phonemics and conditioning
theory are unsatisfactory in the explication of
These, then, are the criteria which have a particular case of interference behavior and
determined the sequence of our presentation: that generative phonology in the sense of
Halle, Chomsky, et al., shows more promise in
1. Hierarchy of difficulty this area.
2. Functional load
3. Potential mishearing The substitution of different sounds for the
4. Pattern congruity interdental fricatives of English by learners
from different L1 backgrounds has been
Matching these criteria against one another is marked by many investigators. Weinreich
not easy task, and there is clearly no single (1966, p. 20) notes that the majority of
‘right’ or ‘best’ sequence of presentation. Our French speakers substitute /s/ and /z/ for
own procedure has been, in general, to put English /T/ and /D/respectively, whereas
those things first that were most important in Russian speakers substitute /t/ and /d/.
the task of communication, either because Berger (1951, pp. 47-51) reports the same
mishandling of them could easily result in substitution for Russian speakers as well as
misunderstanding or because they carried a the substitution of /s/ and /z/ among schooled
heavy functional load and would therefore be French speakers -/t/ and /d/ among
especially obvious and frequent sources of unschooled. Lado (1957) finds /s/ for English
accent. In order to get similar problems /T/ in Japanese speakers, /t/ in speakers of
together, however, we have violated the Thai and Tagalog. Kohmoto (1965) also
mixed criteria of importance. Our preferred reports /s/ and /z/ for /T/ and /D/ in Japanese
pedagogical sequence is: speakers. Angus (1973) reports that Turkish
speakers fluctuate between /t/ and /s/ for /T/.
1. Basic intonation features and patterns The present discussion will be restricted
(including stress, pitc, juncture and primarily to the treatment of Russian based
rhythm) substitution of /t/ for /T/ and Japanese-based
2. Weak stressed vowels substitution of /s/ for the same sound.
3. Strong stressed vowels and diphthongs
4. Voiced stop-spirants It is generally conceded that one kind of
5. Vibrants and liquids interference behavior, phone substitution,
6. Voiceless stops results when a learner unconsciously identifies
7. Spirants or categorizes an L2 sound as being the same
8. Nasals and palatals as a particular L1 sound (even though it
9. Semivowels differs from the L1 sound in the perceptions of
10. Consonant clusters native speakers of the L2) and substitutes the
11. Other intonation features and patterns. latter sound for the former in L2 utterances.
The two questions that must be answered by
an explanation of phone substitution are: (1)
On the Explanation of Phonic On the basis of what property of the L2 sound
Interference does the learner identify the L2 sound –i.e.,
what properties are identified by the learner
William C. Ritchie as being shared by the L2 sound and the
substituted L1 sound? (2) Why does the
That the goal of a foreign language (L2) learner identify the L2 sound on the basis of
course is the modification of the learner and these properties rather than others?
his behavior in some way is beyond dispute. A
major factor in such modification is the A significant explanation of interference must
elimination of the influence of the native be based on a phonological analysis which is
32
justified independently of the specific goal of For the most part, on /t/ and /d/. Van Teslaar
explaining interference. It is possible to (1966) has noted that learners who
construct a phonological analysis specifically pronounce well in a learning situation may
for the purpose of explaining interference revert to interference behavior under the
behavior; but such an analysis would explain strain of conversational conditions. In general
nothing, since it would be entirely ad hoc. If we can expect the learner’s L1 to influence his
our explanation of interference is to be performance more deeply under the
significant, the dimensions we choose in conditions found in conversation than under
identifying or describing L1 and L2 sounds those in a learning situation where the learner
(that is, the answer which we provide for may be allowed to concentrate on the careful,
question 1 in a given case of interference) correct articulation or comprehension of
must be motivated within the analyses isolated sounds or sound sequences. For these
themselves. For example, we may loosely reasons, the study of interference in
describe the motivation behind a conventional conversational performance is likely to be
phonemic analysis as the desire to provide an more revealing than that of interference in
economical description of contrasting classes learning performance. An additional reason for
of phones (each phone described in studying and attempting to explain
articulatory terms), and therefore the conversational rather than learning
dimensions chosen for the conventional performance is the obvious practical one that
description of a sound pattern are those conversational performance is precisely what
articulatory dimensions (and only those) along we wish a course in an L1 to modify –a course
which all members of one class are which does not succeed in the specific task of
distinguished from all members of each other modifying conversational performance must
class. Conventional phonemics, then, provides be considered a failure. What is to be
such dimensions as manner and point of explained, then, is the learner’s performance
articulation, voicedness in the case of in conversation.
consonants, and height and degree of
frontness-backness in the case of vowels; if Conventional phonemics and contrastive
we are to explain phonic interference in terms analysis.
of conventional phonemics, we must answer
question 1 in terms of these dimensions. Most attempts to explain interference in
general have been couched in terms of
A plausible answer to question 2 would be contrastive analysis based on conventional or
that the learner identifies the L2 sound on the classical phonemics. It is thus important to
basis of those of its properties which are ascertain the answers conventional phonemics
distinctive or phonemic on the L1, although can provide for the questions formulated
the obvious subsidiary question arises: On the earlier.
basis of which of its distinctive features is the
sound identified? A phonological analysis of a Phonetic properties in conventional phonemic
specific language (an the general theory of analysis
phonology from which the specific analysis
derives) can be considered as an appropriate Although the strictly articulatory or
basis for the explanation of interference physiological description of speech sound
behavior if (a) it attributes distinctiveness to involves, from a narrowly linguistic point of
that property upon which the learner who view, an arbitrary system of classification, it
speaks the specific language in question bases has been found that the sound patterns of
his identification on the L2 sound or (b) in the languages can be described in terms of a
cases where the learner chooses one from a limited number of dimensions, usually
set of distinctive properties, the analysis expressed in articulatory terminology (as in
provides grounds for explaining this choice, Bloomfield 1933, chap. 6 on “Practical
(for example, on the basis that some Phonetics”; Jakobson, Halle, and Fant 1952;
distinctive properties are more important than and de Saussure 1959, pp. 38-64 on
others in the categorization of sound). “Phonologie”). Basic dimensions are (1)
consonantal versus vocalic, (2) point of
In the initial stages of L2 acquisition a learner articulation among consonants, and frontness-
may fluctuate considerably in the L1 segment backness among vowels, (3) manner of
he substitutes for a given L2 sound. Berger articulation among consonants, and height
(1951, p.47) reports that Russian speakers among vowel, and (4) voiced versus voiceless
learning English substitute /ds/, /tT/, /dD/, among consonants. Since we will be dealing
/s/, /z/ for the interdentals before they settle. only with voiceless consonants here, we can
33
ignore the consonantal-vocalic and voiced- We might seek to explain the learners’
voiceless dimensions. In the consonant behavior in terms of their respective histories
system, the dimensions point and manner of of reinforcement. However, the form of
articulation have several well-known values behavior which must have been reinforced in
(“bilabial,” “dental,” “alveolar,” etc., for the the learner in order for him to exhibit the
point dimension and “stop,” “spirant,” “nasal,” observed interference behavior –that is,
etc., for the manner of dimension). The production of /s/ and /t/ in echoic response to
presence in a given segment of one of these /T/- is highly improbable since it would
values on each dimension implies the absence require a situation in which Russian- and
from that segment of all others on that Japanese-speaking adults produce /T/ and
dimension so that, for our purposes, a require their children to imitate them with /t/
voiceless consonant segment is fully and /s/, respectively.
determined within the sound pattern by its
manner and place of articulation. It is possible that a Japanese adult who has a
lisp history might identify English /T/ with his
Assuming that the usual designations “stop,” earlier attempts to produce /s/ and therefore
“alveolar,” have universal validity –i.e., that substitute specifically /s/ and /T/ but the
these values have the same meaning from acceptance of this as a general explanation is
one phonemic description to another- we have excluded on obvious grounds.
some basis for comparison among sound
patterns. In these terms the variants of The hypothesis that Japanese and Russian
English /T/ and the variants of Japanese /s/ children must in general be trained to
share the value “spirant” on the dimension substitute /t/ for earlier /T/ is not in keeping
“manner of articulation” and differ on the with what is known about child acquisition of
dimension of “point of articulation” in that /T/ phonology: in Lewis’s compilation of 310
is interdental and Japanese /s/ is alveolar cases of phone substitution in French-,
(Bloch 1950, p. 343) Japanese /t/ is dental German-, and English-speaking children there
and therefore “phonetically closer” to /T/ than are no cases of the substitution of the
is /s/ with respect to the point of articulation interdentals for other segments (Lewis 1951,
dimension although, of course, it differs from pp.310-31).
/T/ on the manner dimension in being a stop
rather than a spirant. The characterization of a sound pattern as a
three- or four-dimensional matrix in
Trofimov and Jones (1923, p. 96) describe conventional phonemics is apparently
“normal” Russian /t/ as voiceless dental motivated on the grounds that this
plosive although one of its chief subsidiary arrangement is convenient either for
members is alveolar. Russian /s/ is described organizing field-work or for publication
by the same authors (p.) as a breathed blade- purposes. While the categories that arise from
alveolar fricative. this motivation may offer the investigator a
useful framework, they do not necessarily
Except that Russian /t/ has an alveolar match the way in which the speaker-hearer
allophone and Japanese does not, the variants tacitly categorizes the same segments. In
of the dental stops and alveolar fricatives do order to be relevant to an investigation of
not differ basically between Russian and interference behavior, a linguistic description
Japanese. must make the claim that those categories
which it posits are, in fact, the categories in
Conventional explanation of substitution for terms of which a native speaker-hearer of the
English /T/ language categorizes or interprets speech
utterances. Whether or not the native
Phonetic considerations speaker-hearer unconsciously categorizes,
e.g., consonantal sounds in accordance with
As noted above, the articulatory properties of their point and manner of articulation or in
the allophones of Russian /t/ and those of terms of some other set of dimensions and
Japanese /t/ are quite similar, as are those of values, is an empirical question and a very
Russian /s/ and Japanese /s/. it seems basic one for the explanation of interference
improbable, then, that the substitutions of behavior.
different sounds for English /T/ by Russian
and Japanese speakers can be explained on Distinctiveness in conventional phonemics
purely articulatory grounds.

34
Although conventional phonemicists have not relevant to the explanation of interference
always agreed in detail among themselves as behavior.
to the basis for phonological analysis, the
crucial distinction in phonemic analysis is Phonological properties in a generative
clearly that between contrastive and non- phonology
contrastive distribution of phonetically similar
segments. For example, Bloch (1948) finds There are two sets of dimensions or features
the set of dental stops in Japanese to be in in a generative phonology: (1) classificatory
contrast with (and therefore phonemically features, which are two-valued, and (2)
distinct from) the set of a dental (or denti- phonetic features, which may have more than
alveolar) affricates on the grounds of such two values (Chomsky 1964). The first set is a
pairs as /mats.to/ ‘if one waits’ and /mat.te/ modification of the Jakobsonian features. It
‘waiting’. serves to categorize segment types in the
underlying representations of morphemes
Bloch finds, on the basis of conventional from which the phonetic representations (in
criteria, that the phonetic difference between terms of phonetic features) are derived by the
the dental and alveolar point of articulation is rules of the phonology (Chomsky 1964; Halle
not distinctive but it is predictable on the 1964a, 1964b; McCawley 1965). The
basis of manner of articulation –stops are underlying representations of morphemes,
dental, spirants alveolar. It might be then, are matrices with segments as columns
hypothesized that the possible substitution in and features as rows. Except for certain cases
Japanese speakers of /t/ for /T/ does not which will be noted immediately, each
actually occur because the basis for such a segment is designated within the matrix as
substitution –that is, the greater proximity of having a value with respect to a given
/t/ to /T/ than of /s/ to /T/ with respect to classificatory feature. However, if the
point of articulation- is undermined by the designation of the value of a particular
lack of contrast between dental and alveolar segment with respect to a particular feature is
point of articulation in Japanese. predictable by the rules of the phonology,
either from the values of other features in the
However, the same explanation does not hold segment or from the values of features in
for Russian. Apparently, the same relationship neighboring segments, then that feature
between dental stop and alveolar fricative designation is left unspecified in the
holds in Russian (i.e., dental versus alveolar underlying form of the morpheme. Such
point of articulation is non-distinctive) since designations will be supplied by the rules. For
Russian /t/ has alveolar allophones. Thus, example, McCawley (1965) finds that the
according to the hypothesis given above, we affricateness and length of /ts/ in, e.g., the
would expect the Russian, like the Japanese, Japanese form /mats.to/ ‘if one waits’ is
to substitute /s/ for /T/ whereas he actually predictable by two general rules. The first
substitutes /t/. (rule 25, p. 136) states that when u occurs
between two voiceless obstruents in
Although the above treatment of interference underlying representations, it is presented
behavior in terms of conventional phonemics phonetically by its voiceless counterpart /U/
does not exhaust the possibilities, a and the second (rule 26, p. 137) that all
satisfactory explanation of interference in dental stops that precede nonconsonantal,
these terms is difficult, if not impossible. diffuse, grave segments (including /U/) are
phonetically affricate. (Apparently, Bloch
Generative phonology interpreted McCawley’s phonetic sequence
/tsU/ as /ts/) Thus affrication need not to be
A generative phonology, as a part of a full represented in the underlying forms of
generative grammar, describes an aspect of morphemes containing phonetic affricates
the speaker-hearer’s linguistic competence. before underlying U since this feature will be
That is, an empirically adequate generative supplied by the rules of the grammar.
phonology characterizes that information upon
which the native speaker-hearer’s Part of the problem of explaining a particular
categorization or interpretation of speech instance of phone substitution is establishing
sounds and sound sequences is based (though what interpretation the learner has imposed
its relation to actual categorization on the context in which the substitution
performance may be quite indirect). In other occurs. This task is a highly complex one and
words, it makes precisely the claim that a we shall not attempt to perform it for the
linguistic description must make if it is to be particular case of interference under
35
discussion here. Instead we will limit the basis of the values which that segment alone
domain of our explanation of substitution for has with respect to other features. For
English interdentals to a phonetic environment example, the fact that Japanese [s] is strident
which can be assumed to have a minimal (rather than mellow) is predictable from the
contextual influence on the learner’s fact that it is “distinctively” obstruent, grave,
interpretation of the consonants in question. continuous, and nonsharp (rule 23, McCawley
1965, p.136).
Pause is, perforce, always identifiable by the
learner as a boundary in L2 utterances; we Halle (1959) imposes on the inventory of
assume that true vowels in L2 utterances are underlying sentences segments the condition
more easily identified as such than, say, that the maximum number of feature
glides are as glides; initial consonant clusters, specifications is underlying segments be
if they exist in the L1 at all, are likely to rendered predictable by phonological rule. He
exhibit interdependencies among their states (p.34) that this condition is equivalent
constituent segments which may influence the to the requirement that the inventory of
learner’s identification of initial clusters in L2 segments be determined or described by a
utterances. With these factors in mind, we decision tree (more specifically the simplest
choose to limit our explanation to decision tree) in which each node represents a
substitutions in the position between pause feature and each branch from a node
and a true vowel. represents a value (+ or -) of the feature. The
first (top) node divides all segments into two
Assuming a direct relationship between the classes (those which are [+consonantal] and
substantive universal classificatory features those which are [- consonantal]), the second
(that is, ·stridency,” “continuity,” node divides each of these further into two
“compactness,” etc., e.g. Halle 1964a) and classes ([+ vocalic] and [- vocalic]) and so on.
their phonetic correlates, we may evaluate Each path through the tree represents a
English /T/, Japanese /t/ and /s/, and Russian distinct segment. That is, each segment is
/t/ and /s/ as consonantal, nonvocalic, diffuse identified by answering a sequence of
(versus compact), acute (versus grave), and questions about it –is it consonantal? Is it
voiceless Japanese and Russian /t/ are vocalic? Diffuse?, etc. However, the process of
discontinuous and mellow; /s/ in both identification of any one segment is generally
languages is continuous and strident. The more efficient for a given language if the
facts to be explained, then, are that the questions are asked in one order than if they
Russian speaker categorizes /T/ as primarily are asked in another. Thus, as a consequence
mellow /as like /t/) whereas the Japanese of representing the structure of the segment
categorizes it as primarily continuous (like his inventory as the simplest decision tree, a
/s/). hierarchy is established among the features.
Halle writes (1959, p.34): “The hierarchy of
Explanation in terms of generative phonology features seems to provide an explanation for
the intuition that not all of the features are
Distinctiveness in a generative phonology equally central to a given phonological
system.”
In a generative phonology a property of a
particular segment may be said to be Although a generative phonology makes no
distinctive or phonemic in that segment if it is direct claims about the perception of
not predictable by a phonological rule. If it is utterances, we might hypothesize a rather
predictable then it is nondistinctive. Bloch simple relationship between the phonological
found the segment sequences [ts] and [t] to code and speech perception with respect to
be in contrast with the basis of such forms as centrality of features within a system. This is
[mats.to] ‘if one waits’ and [met.te] ‘waiting’. namely that the information represented by
In terms of a generative phonology, on the the feature hierarchy on the decision tree is,
other hand, these two segment sequences are all things being equal, reflected in perception
not distinct since the affricateness (or, in by a ‘hierarchy of cue preference’ (Bruner,
Jakobsonian terms, the stridency) of [ts] is Goodnow, and Austin 1956, pp.31, 35). The
predictable. In this example, the value phonetic correlates of a feature which is high
strident (versus mellow) of the segment [ts] in the phonological decision tree will have
is predictable from the segment’s position greater importance in perception or, to use
before u in underlying representations. The the term of Bruner et al. (p. 31), a higher
value of a particular segment with respect to a ‘degree of criteriality’ in the classification of
given feature may also be predictable on the
36
speech sounds by native speakers than that of phoneme inventories for our explication of
a lower-placed feature. interference in these cases.

Explanation From what has preceded, it should be clear

that an explanation of phone substitution for
The value of any segment with respect to the /T/ of the sort offered above for Japanese and
stridency feature is predictable in Japanese Russian speakers is based on the
(morpheme-structure rule 7, p. 129); independence of the stridency and continuity
phonological rules 23 and 26, pp. 136- features. If a language has only strident
37[McCawley 1965]). On the other hand, the continuants and mellow stops among its
value of continuity is predictable only in very obstruents then we cannot explain the
limited contexts. Although McCawley does not substitution of, e.g., /t/ for /T/ by speakers of
impose Halle’s simplicity criterion on the that language on the basis of the higher
inventory of underlying segments in his position of stridency (over continuity) in the
analysis of Japanese, the complete hierarchy of features since stridency and
predictability of stridency values and the continuity are combined in the same feature.
incomplete predictability of continuity values French (Jakobson and Lotz 1949), Turkish
would be represented in tree-diagram form by (Lees 1960), Tagalog (Bloomfield 1917), and
the placement of continuity above stridency in Thai (Abramson 1962) all have coalesced
the feature hierarchy. We would thus predict continuity and stridency. Turkish and French
that a Japanese speaker, all things being speakers exhibit fluctuations between /s/ and
equal, will ‘attend to’ the phonetic correlates /t/ as might be expected on the grounds of
of continuity in speech utterances and ignore coalescence of stridency and continuity. On
the correlates of stridency. This would explain the other hand, the explanation of Thai- and
his production of /s/ for /T/ since these two Tagalog-based substitution of /t/ for /T/ is not
segments share the same continuity value possible in this way. We might conjecture
though they differ with respect to stridency. that, although there is no independent formal
reason to differentiate between continuity and
The Russian situation presents a slightly more stridency in these languages and no way to
subtle problem since both continuity and establish whether the distinctive phonetic
stridency are distinctive. However, the correlates upon which categorization of
description of the sound pattern of Russian sounds is based are those of stridency and
has a whole is simpler if the stridency feature those of continuity, the feature in question is,
is placed above the continuity feature in the in fact, stridency, since this would explain the
underlying decision tree than if the order is categorization of mellow /T/ with mellow /t/
reversed (Halle 1959, p. 46). The Russian by speakers of these languages. However, in
speaker will thus attach a higher degree of the absence of independent evidence for this
criteriality to the stridency dimension than to conjecture the explanation is ad hoc.
the continuity dimension. He thus groups /T/
with /t/ on the basis of shared mellowness. A Pedagogical implications
general rule in Russian states that all mellow
non-nasal consonants are stops (Halle 1959, Phonological systems are notoriously well
rule p 5a, p. 65). Having categorized /T/ as entrenched in adults. Halle (1964b, p. 344)
mellow, the Russian speaker derives the conjectures that “… changes in later life are
information that it is also a stop and produces restricted to the addition of a few rules in the
/t/ instead. grammar and that the elimination of rules and
hence a wholesale restructuring of his
Explanations of the other cases of phone grammar is beyond the capabilities on the
substitution enumerated above may be average adult.” Although this conjecture is
attempted with the reservation that the made specifically with respect to changes in
structure of the underlying segment inventory the native-language grammar of an adult its
of a language cannot be known with any implication for L2 learning are clear; in fact, if
degree of certainty without a set of explicit the conjecture is extended to changes in the
rules which relate underlying forms of their conceptual structures underlying speech
consequences –that is, without a generative perception in general (including the perception
phonology. Since treatments of Thai, Tagalog, of L2 utterances by learners) then, as is well
and French from this point of view are not known, considerable evidence can be adduced
available (Lees [1960] provides a generative from the study of L2 learning to support it. In
phonology of Turkish), we must rely on some cases, though, learners do gain a strong
distinctive feature analyses of conventional intuition for an L2 and it is well to inquire how
37
we may increase the probability that a course similarly with stridency, continuity, etc. this
in an L2 will produce such learners. method has several drawbacks. First of all, it
is virtually certain to be more time consuming
It would be clear that the observed cases of than its effects warrant. Second, it may be
interference in Japanese- and Russian-based confusing for the learner to find (at least in
articulations of English interdentals cannot be this way) that one categorization places two
explained simply as the failure of the Russian given segments in different classes and
speaker to ‘discriminate the stimuli /T/ and another categorization places the same two
/t/’ or of the Japanese speaker to segments in the same class. It is conceivable
‘differentiate the responses /T/ and /s/’.’ that, if features and their categories are
These problems apparently lie in the presented in this way, the learner will resort
identification of an unfamiliar event-type (the to the completely meaningless and ad hoc
sound /T/ in English utterances) in terms of a memorization of features and the segment
highly-structured cognitive system (the categories that they determine.
phonology of the L1) which is not appropriate
to the task of identifying the event-type in Phonetic features (as opposed to classifactory
question. The solution to these problems lies features) represent the intrinsic physical
not in the modification of the superficial, properties of sounds; the categories which
particular consequences of the underlying they determine may therefore be termed
general system, but in a basic alteration of formal in the sense of Bruner et al. (1956, pp.
the underlying general system itself. 5-6). The method of teaching classificatory
features and categories suggested earlier is
One way in which we might proceed to alter based on the supposition that classificatory
the linguistic cognitions of learners of English categories might fruitfully be learned formally
as an L2 is by assigning the task of learning to in terms of the intrinsic physical properties of
read aloud systematic phonemic (or perhaps the sounds which correlate with the segments
more abstract) representations of English categorized. However, the role of
utterances. Since the phonology of a language classificatory features and categories in a
is a set of rules which relate syntactic phonology is functional or relational rather
representations of utterances to their phonetic than formal in that they represent the
realizations, the ability to ‘read’ the syntactic relationships among segments in the sound
representation of an utterance can be pattern and the way in which segments enter
considered as equivalent to a tacit knowledge into the applicability of phonological rules or
of the phonology of the language. The tacit the way in which segments ‘pattern.’ These
application of phonological rules in the reading facts suggest that the representation of
of an abstract transcription demands that particular segments in terms of classificatory
segment-letters be categorized by the reader features might best be learned simply as one
in accordance with the classificatory feature aspect of learning the rules in which they
complexes which characterize their appear. For example, the choice of the
corresponding segments (since the rules are phonemic form of the regular plural and
formulated in terms of these features). For possessive of nouns, and the third person
this reason, a major goal of phonological singular of verbs depends upon categorizing
instruction is the learner’s acquisition of the correctly the final segments of noun and verb
ability to categorize segments accordingly. stems- first with respect to gravity and
stridency (since stems with final non-grave,
Various simple techniques for accomplishing strident consonants take the form /+z/), and
this task come to mind: for example, in order second with respect to voiced-voiceless (since,
to teach the consonant system, we might of the stems that do not come under the
simply present the learner with single-syllable above rule, those that end in voiced segments
utterances composed of a consonant of the L2 take /z/ and those that end in voiceless
followed by the optimal vowel /a/ and ask him segments take /s/). The Japanese speaker’s
to assign the syllables to categories, problems distinguishing /T/ from /s/ may thus
reinforcing him positively when he groups be subsumed under the general problems of
them in accordance with the compactness first, distinguishing strident continuants from
value of the initial consonant and negatively mellow continuants and second, choosing the
when he does not. Then present him with the correct ending for regular noun plurals and
same syllables (or perhaps only with syllables possessives, and for third singular verbs. In
that have consonants of the same this case the acquisition of the ability to
compactness value) and ask him to categorize categorize segments correctly with respect to
them in accordance with their gravity values; the stridency feature takes on a functional
38
significance which is lacking in the learning Selinker 1972). There are numerous
procedure suggested earlier. In fact, the precedents for using this type of data. Speech
ability to form correctly novel regular noun errors committed in the native language have
plurals and possessives, etc., under provided some of the most useful evidence for
conversational conditions is strong (if not formulating hypotheses regarding the nature
conclusive) evidence that the learner has of the mechanisms underlying speech
acquired the classificatory categories of production (Lashey 1951; Hockett 1967;
strident and mellow regardless of whether the Boomer and Laver 1968; Laver 1969; MacKay
learner ‘differentiates the responses’ /T/ and 1970; Nooteboom 1969), the psychological
/s/. reality of theoretical linguistic concepts
(Fromkin 1971) and the perceptual strategies
Similarly, we might take the Russian’s brought to bear in the comprehension of
mastery of the continuant-discontinuant linguistically complex structures (Blumenthal
distinction a part of his acquisition of the rule 1967; Bever 1970).
(noted by Sapir, 1925) that certain noun
stems which end in voiceless continuants have Obviously, no single mechanism can be the
corresponding stem-final voiced continuants in source of all speech errors made by second
their pluralization (for example, /nayf/- language learners. A number of causes have
/nayvz/, /bQT/-/bQDz/, /haws/-/hawz+z/, etc. been recognized: interference formative
This rule will in no case apply to nouns with language, the application of general learning
stem-final stops. Thus we might expect a new strategies similar to those manifested in first
formation /feyDz/ (meaning ‘religious language acquisition, such as
dominations’) as the plural of /feyT/ or overgeneralization of linguistic rules (Richards
*/kÃbz/ related to /kÃp/. in learning this rule, 1971, 1975) the context of the learning
then, the Russian speaker must learn to intuit situations (cf. Selinker’s /1972/ ‘transfer of
the systematic distinction between training’ errors), and emotional factors
continuants and stops and, as a consequence, connected with the pressure of
the functional distinction between /T/ and /t/. communicating, e.g. the ‘strategies of
communication’ suggested by Selinker. In
It should be clear that the remarks above are addition, it has been suggested that certain
only suggestive and that the construction of a untestable ‘avoidance strategies,’ (cf.
maximally effective course in the practical Schachter 1974) shape the learner’s output.
phonology of an L2 is an intricate task into
which all kinds of factors enter. I do not claim The second of the casual mechanisms
that Japanese speakers will suddenly mentioned above –general learning strategies
distinguish /T/ from /s/ upon learning to similar to the processes at work in native-
pluralize nouns –only that systematic factors language acquisition- has enjoyed the
are of basic importance in the construction of greatest amount of attention in literature on
courses in L2’s and that we cannot hope to second language learning. Unfortunately,
maximize the effectiveness of L2-phonology much of the evidence cited in support of this
instruction without giving them central mechanism is far from unequivocal, which
consideration. raises the disturbing possibility that native-
like learning strategies ay be invoked to
Toward a Psychological Theory of account for any error which is not a totally
Interference in Second Language unambiguous example of one of the other
Learning processes. This paper is concerned with the
most widely documented source of error in
J Ronayne Cowan adult second language learning: the intrusion
of the native language. The discussion will be
During the past decade, the major thrust of confined to the type of interference one
psycholinguistic research has been the frequently encounters in the speech of adult
investigation of how children acquire their language learners: the application of a
native language. Now it would appear that grammatical relation to the native language to
researchers are gradually becoming interested the morphemes in the second language (cf.
in exploring the cognitive process recruited by Weinreinch 1953, p. 30). As a first step
adults engaged in learning second languages. toward the formulation of a psychological
The data relevant to researching this topic are theory of interlingual interference, I will
the errors made by the learner in his attempts propose two principles which predict the
at meaningful communication in the second occurrence of production errors, and illustrate
language (Weinreich 1953; Corder 1967; how they account for semantic errors which
39
arise as a result of the learner’s false equation In seeking to formulate psycholinguistic
of a native language structure with a structure principles which satisfy the above-
in the target language. Next, I will compare mentioned requirements for a theory of
this psychological explanation of one kind of syntactic interference, we must proceed
interlingual interference with an alternative from the assumption that speech is rule-
account which utilizes linguistic universals as governed behavior, and these rules are
an explanatory device. The purpose of this reflected in cognition. This being the case,
comparison is to draw a distinction between the most natural and general principle for
causal and noncausal explanations. specifying the process by which a
syntactic construction in the native
To be of any interest, a theory of interlingual language is seen as analogous to a
interference must satisfy two requirements. syntactic construction in the target
First, it must provide an explicit specification language is shown in 1:
of the criteria by which the learner pairs
elements of the native language – 1. Principle 1: if the learner views the output
phonological, lexical, syntactic, semantic, or of a rule x in the second language to be
rhetorical- with elements in the target functionally equivalent to the output
language. Failure to satisfy this requirement structure of a rule y in the native
markedly reduces the explanatory power of a language, then rule y will tend to be
theory, since the underlying causes of applied in contexts where the learner
interference are then excluded from empirical deems x structures appropriate.
investigation. Interlingual interference is
generally understood to mean the systematic I will define a functional equivalence as
influence of the native language on the the recognition that the two utterances
learner’s attempts to use (produce and are equatable in terms of their
perceive) the target language. This most illocutionary function, i.e., that they both
frequently takes the form of native language negate, question, modify, express a
intrusions, but whether or not these intrusions specific proposition such as existence,
are direct importations of native language predication, etc., or a combination of
elements, or whether they take on some form these functions.
that lies between the native language and the
target language is irrelevant. The fact remains The principle formulated in 2 states the
that unless interference is to be viewed as optimal conditions for the occurrence of
totally unsystematic, specific conditions which interference due to applying the native
lead to the learner’s associating native language rules.
language elements with corresponding
elements in the native language which will 2. Principle 2: When the learner employs
come under the influence of the former must Principle 1, the maximum possibility of
exist. It seems intuitively obvious that the errors occurring exists when the formal
bases for establishing correspondences could properties of rules x and y are antithetical.
differ depending on the nature of the
elements. Thus, the grounds for phonological By formal properties, I mean the set of
interference may be acoustic and/or conditions imposed upon the manner in
articulatory, whereas lexical interference which a rule operates. The following
might result from pairing that has an acoustic example demonstrates that these
or graphic basis. There is, however, no reason conditions reflect the mental reality of the
to exclude the possibility that a single rule for the speaker, and that they
principle for establishing correspondence could frequently shape the strategies he
apply to a number of different kinds of employs for communicating in the second
interference. language.

The second requirement that a theory of Both English and Hausa, a Chadic
interlingual interference must satisfy is that it language spoken primarily in Nigeria, but
contain principles which predict the likehood also as a lingua franca along the coast of
of an error occurring once a correspondence West Africa, possess positive and negative
has been established. These principles should questions of the type shown in 3.
be formulated with sufficient precision to
permit their extension to new cases in 3. a) is your father here? (positive question)
different languages.

40
 b) Isn’t your father here? (Negative then the formal differences between the
question two systems can be symbolized as follows:

However, Hausa, like Japanese, Hindi, and 5. a) QE,Æ QE = UpS

other languages, employs a different set b) QH = UpS
of answers for negative questions than c) Æ QH = UÆpS
English. The alternatives for English and
Hausa are shown in 4. Since negative questions have the same
basic functional range in both languages,
4. English i.e., they are both used to request
Q. Isn’t your father here? information, native speakers of English, in
A: a) No. (He is not here). accordance with Principle 1, set up the
b) Yes. (He is, in fact, here) equation ÆQE = ÆQH when learning
Hausa Hausa. (I am purposely ignoring the more
Q: Babu tsohonka a nan? restricted functions of English negative
(Is not father you here?) questions –that of expressing annoyance,
A: c) I, babu (Yes, he is not (here).) disagreement, e.g., “Aren’t you ashamed
d) A? a, akwai. (No, he is (here).) of yourself?” = “You really ought to be
ashamed of yourself!” or disbelief: “Didn’t
Pope (1973) has argued that 3 and 4 someone call last night?” = “Oh, come on!
provide evidence for two types of Admit that someone called last night!”
question-answering systems. She since this discussion is confined to the
designates the system in English a more common use of negative questions –
positive-negative system, whereby an requesting information- these other uses
answer is negative if it contains sentential are not germane. Note also that, when
negation in the highest clause, and used in these restricted capacities, English
positive if it does not. Hence a no answer negative questions have distinctly different
to 3a and 4a expresses a negative intonation contours.) However, 5 shows
disagreement and agreement respectively; that the formal properties of the rules in
and a yes answer to the same two each language are diametrically opposed.
questions expresses positive agreement The result, predictable from principle 2, is
and disagreement, respectively. In maximal interference, whereby the learner
contrast, Hausa is seen as an agreement- employs his native language answers in
disagreement system: an answer agrees if response to the Hausa questions. Learners
it matches the question with respect to can only correct this error through
negativity and disagrees if it does not. understanding the mental image of the
Thus, answer 4c is an expression of Hausa structure. He must grasp the notion
negative agreement, and 4d expresses that the negative question in 6a has the
positive disagreement. formal properties shown in 6b.

Pope’s distinction obscures a basic 6. a) Wannan ba zabo ba né? (that neg.

generalization about yes-no questions Guinea fowl neg. Is “Isn’t that a guinea
which is apparently true for all languages: fowl?”)
that they request an affirmation or denial b) Is it the case (p = that isn’t a guinea
of the truth value of the proposition being fowl)?
questioned. As a result, her description
fails to capture the psychological reality of It is worth noting this example constitutes one
the linguistic structures in both languages. of the rare cases where one could safely
In fact, the only difference between predict that second language learning mirrors,
English yes-no questions and their Hausa to some extent, the sequence that takes place
counterparts is that both negative and in child language acquisition. This follows
positive yes-no in English request an because the Hausa child has a more
affirmation or denial of a positive semantically complex system than the
proposition. This condition also holds for English-speaking child. (For a discussion of
Hausa positive yes-no questions; however, how one might intuitively define complexity in
negative yes-no questions in Hausa child language, and how it may affect
request affirmation or denial of a negative acquisition, see Slobin 1973, p. 188) Since
proposition. If we let QE and QH represent positive questions are more frequent than
positive questions in English and Hausa, negative questions, we would expect the child
respectively and p stand for proposition, to master the equation in 5B first. Confronted
41
with examples of negative questions, he will cognate. It is difficult to see how one could
first employ the answers to positive questions relate cognate forms to linguistic universals,
and only later acquire the additional system but even if one were to construct some
shown in 5c. Bilingual children may initially universal or typological hierarchy where
have great difficulty with the two systems. A languages with no cognates are more highly
study of the development of yes-no questions ranked (less marked) than languages which
in English-speaking, Japanese-speaking, and contain many cognate forms, one would still
English-Japanese bilingual children by be left with the task of explaining why
Akiyama (1977) supports these hypothesis. cognates are both a source of positive,
facilitative transfer as well as negative,
An alternative explanation of the semantic retroactive interference. What is needed is an
errors which arise when English speakers explanation of the mechanism underlying
attempt to learn the Hausa yes-no question these two concepts.
answering system is provided by Eckman’s
(1977) “markedness differential hypothesis.” Interestingly enough, just such an explanation
Under this proposal, the learner’s errors are was discovered a little over thirty years ago.
accounted for and predicted by making In summarizing the experimental research in
reference to a typology of markedness. transfer and retroactive interference in verbal
According to Eckman, those areas of the learning, Osgood (1949) developed a three-
target language which differ from the native dimensional model which has been confirmed
language and are more marked in terms of by several tests, e.g., Bugelski and
linguistic universals will be more difficult, Cadwallader (1965), Dallet (1962), and Wimer
giving rise to predictable errors. However, (1964). In Osgood’s model, when the stimuli
those areas of the target language which are are similar but the responses different,
different from the native language but are interference invariably occurs. If, on the other
more marked will not prove difficult. To apply hand, both responses and stimuli are the
the markedness differential hypothesis to our same, positive transfer occurs. In the neutral
previous example, we must first assume the case, where neither transfer nor interference
existence of a universal hierarchy of yes-no takes place, the stimuli are different (Young
questions where, using Pope’s designations, and Underwood 1954). Osgood’s model
agreement-disagreement systems are more provides an explanation of the mechanisms
marked than positive-negative systems. The underlying lexical transfer and interference in
difference between English and Hausa yes-no second language learning. Whenever the
question systems is thus reflected in the learner pairs two words that are highly similar
universal hierarchy and allows only one in both languages, e.g., French coin, ‘corner’,
possible prediction, i.e., errors will arise when and English coin, but have different meanings,
English speakers attempt to learn how to we have a case of interference due to false
manipulate the more marked system, Hausa, cognates. Conversely, true cognates will
but not vice versa. Note also that the facilitate learning, since they have the same
markedness differential hypotheses can also meaning. Noncognate items cannot be paired
lay claim to the aforementioned prediction since no basis, i.e., similarity, for this exists,
about the development of yes-no questions in so neither transfer nor interference effects will
the speech of Hausa children by simply stating be observed.
that the less-marked system is acquired first.
The two principles developed earlier represent
How are we to decide which of these two an attempt to refine the psychological concept
explanations, both of which make correct ‘interference’, in such a way that it may be
predictions concerning the example under meaningfully applied to explain cases of
discussion, is to be preferred? Eckman syntactic and semantic errors in second
conveniently provides grounds for rejecting language learning. My claim is that these
the explanatory power of the markedness principles have far more explanatory power
differential hypothesis. He states that “it than alternative accounts of interlingual
would be possible to falsify the MDH if it could interference like the markedness differential
be shown that the areas of difficulty that a hypothesis, because they are rooted in the
given language learner has are not those mechanisms of perception and cognition
areas of the target language which are rather than in a classificatory device such as
different from and more marked than the linguistic universals. It is quite erroneous to
native language” (Eckman 1977, p. 327). One assign linguistic universals a causal status in
of the most widely documented causes of the rather cavalier fashion employed by
errors which fulfills this condition is the false Eckman. To maintain that universals are ‘a
42
reflection of human condition’ (Eckman 1978, errors that are accounted for by Principle 2.
p. 329) simply sidesteps the more complex the first of these had to do with reading
issue of how they explain why learners make comprehension errors which occur when
errors. If, for example, we grant that there native speakers of English assign an SVO
are certain universal phonetic tendencies with interpretation to an initial Noun-Verb-Noun
result from the physiology of the human vocal sequence in learning to read German. Because
tract and central nervous system, we are still German frequently employs an antithetical
faced with the task of explaining how these order, OVS, for this surface structure pattern,
universal tendencies constitute a causal incorrect interpretations which involve
explanation of phonological interference in switching the subject and object can occur if
second language learning. Suppose we grammatical clues such as case and subject-
observe that a native speaker of English verb agreement are neutralized. My second
attempting to learn Farsi first produces the example illustrated how reading confusion can
postvelar stop /q/ as a voiced alveolar tap, occur when the language learner relies on a
but later shifts to substituting a /g/ for that native-language strategy embodied by one
phoneme whenever he must produce it. We syntactic process (pronominalization), which is
may observe that the markedness differential opposed by an antithetical process (deletion)
hypothesis correctly predicts that an error will in the second language. Native speakers of
occur, since Farsi /q/ differs from the English come to rely on pronominal signals to
phonemic inventory of English and is also establish basic anaphoric relationships that
more marked, i.e., occupies a lower position enable them to read with comprehension.
in a universal hierarchy of phonetic segments. Hindi, however, establishes these same
But even though an error is correctly anaphoric relationships by deleting the very
predicted, it turns out to be more complex pronominal signals which are so crucial for the
than a single substitution, and nothing English reader. The comprehension
resembling a causal explanation of both errors breakdowns that result when English speakers
has been provided. In fact, the underlying search in vain for the co reference signals
mechanism which results in the English they are used to follows naturally from
speaker’s initial erroneous production must Principle 2. it is interesting to note that
surely involve his pairing the sound he production errors made by English speakers
thought he perceived with the one in his own learning Hindi also testify to the importance of
inventory which seemed most acoustically this opposition in syntactic processes.
similar to it. (A sophisticated language learner Learners frequently insert subject pronouns
with phonetics training might attempt to where the native speaker omits them, thereby
substitute a retroflexed flap, which is even producing an overly-formal style of spoken
closer.) When the true nature of the sound is Hindi.
made clear to the learner, he may
nevertheless not be able to produce it with To summarize, in this paper I have argued
perfect accuracy, at times substituting the that a psychological theory of interlingual
phoneme in his native language which is most interference must satisfy two requirements
similar in terms of articulation, and later some implied by the notion ‘interference’: it must
interlanguage variant that is neither in the specify how elements in the target language
native or in the target language. This are associated with corresponding elements in
example, which is taken from the author’s the native language and stipulate the likehood
personal experience as a language learner, of errors resulting from such associations. Two
clearly shows that the causal mechanisms principles which meet these requirements
underlying phonological interference can only were proposed to account for semantic
be explained by a theory which employs interference errors. A subsequent comparison
comparisons that are part of the analytical of these principles with an alternative account
component incorporated in models of speech of the error phenomena was carried out to
recognition, e.g., Stevens and Halle (1967), a demonstrate the distinction between
production. explanations which are causal and those
which are predictive but noncausal. It was
It is perhaps overly ambitious to expect that a suggested that a causal theory of interference
single set of psychological principles like those would make use of mechanisms characteristic
proposed in this paper could account for of a broader theory of perception and
interlingual perception errors as well as cognition.
production errors, given the difference
between these two processes. Nevertheless,
in Cowan (1976), I cited some perception
43
 Interlanguage that such data would be those behavioral
events which would lead to an understanding
Larry Selinker of the psycholinguistic structures and
processes underlying attempted meaningful
This article discusses some theoretical performance in a second language. The term
preliminaries for researchers concerned with meaningful performance situation will be used
the linguistic aspects of the psychology of here to refer to the situation where an adult
second language learning.1 These theoretical 4
attempts to express meanings, which he may
preliminaries are important because without already have, in a language which he is in the
them it is virtually impossible to decide what process of learning. Since performance of
data are relevant to a psycholinguistic theory drills in a second language classroom is, by
of second language learning. definition, not meaningful performance, it
follows that from a learning perspective, such
It is also important to distinguish between a performance is, in the long run, of minor
teaching perspective and a learning one. As interest. Also, behavior which occurs in
regards the teaching perspective, one might experiments using nonsense syllables fits into
very well write a methodology paper which the same category and for the same reason.
would relate desired output to known input in Thus, data resulting from these latter
a principled way, prescribing what has to be behavioral situations are of doubtful relevancy
done by the learning perspective, one might to meaningful performance situations, and
very well write a paper describing the process thus to a theory of second language learning.
of attempted learning of a second language,
successful or not: teaching, textbooks, and It has long seemed to me that one of our
other external aids would constitute one, but greatest difficulties in establishing a
only one, important set of relevant variables. psychology of second language learning which
In distinguishing the two perspectives,2 claims is relevant to the way people actually learn
about the internal structures and processes of second languages, has been our inability to
the learning organism take on a very identify unambiguously the phenomena we
secondary character in the teaching wish to study. Out of the great
perspective; such claims may not even be conglomeration of second language behavioral
desirable here. But such claims do provide the events, what criteria and constructs should be
raison d’etre for viewing second language used to establish the class of those events
learning from the learning perspective. This which are to count as relevant in theory
paper is written from the learning perspective, construction? One set of these behavioral
regardless of one’s failure or success in the events which has elicited considerable interest
attempted learning of a second language. is the regular reappearance in second
language performance of linguistic
In the learning perspective, what would phenomena which were thought to be
constitute the psychologically relevant data of eradicated in the performance of the learner.
second language learning? 3My own position is A correct understanding of this phenomenon
leads to the postulation of certain theoretical
constructs, many of which have been set up
1
This paper was begun during the 1968-69 academic to deal with other problems in the field. But
year while I was a visitor at the Department of Applied they also help clarify the phenomenon under
Linguistics, University of Edinburgh. Many students and
teachers at Edinburgh and at Washington, through
discussion. These constructs, in turn, give us
their persistent calls for clarity, have helped me to a framework within which we can begin to
crystallize the ideas presented in this paper to isolate the psychologically relevant data of
whatever level of clarity is attained herein… second language learning. The new
2
It is not unfair to say that almost all of the vast perspective which an examination of this
literature attempting to relate psycholinguistics to
phenomenon gives us is thus very helpful both
second language learning, whether produced by
linguists or psychologists, is characterized by confusion
between ‘learning’ a second language and ‘teaching’ a the way to relevant data. See, for example, Fodor
second language (see also Mackey in Jakobits 1970, p. (1968, p. 48) n”…how we count behaviors and what is
ix) This confusion applies as well to almost all available as a description depends in part on what
discussions on the topic one hears. For example, one conceptual equipment our theories provide…”
4
might hear the term ‘psychology of second language Adult is defined as being over the age of twelve. The
teaching’ and not know whether the speaker is notion is derived from Lenneberg (1967, e.g., pp. 156,
referring to what the teacher should do, or both. This 176) who claims that after the onset of puberty, it is
terminological confusion makes one regularly uncertain difficult to master the pronunciation of a second
as to what is being claimed. language since a critical period in brain maturation has
3
The answer to this question is not obvious since it is been passed, and “…language development tends to
well known that theoretical considerations help point ‘freeze’” (Lenneberg 1967, p. 156).
44
in an identification of relevant data and in the time table;5 there is no direct counterpart to
formulation of a psycholinguistic theory of any grammatical concept such as universal
second language learning. The main grammar; there is no guarantee that this
motivation for this article is the belief that it is latent structure will be activated at all; there
particularly in this area that progress can be is no guarantee that the latent structure will
made at this time. be realized into the actual structure of any
natural language (i.e., there is no guarantee
“Interlanguage” and Latent structures that attempted learning will prove successful),
and there is every possibility that an
Relevant behavioral events in a psychology of overlapping exists between this latent
second language learning should be made language acquisition structure and other
identifiable with the aid of theoretical intellectual structures.
constructs which assume the major features
of the psychological structure of an adult The crucial assumption we are making here is
whenever he attempts to understand second that those adults who succeed in learning a
language sentences or to produce them. If, in second language so that they achieve native
a psychology of second language learning, our speaker competence have somehow
goal is explanation of some important aspects reactivated the latent language structure
of this psychological structure, then it seems which Lenneberg describes. This absolute
to me that we are concerned in large part with success in a second language affects, as we
how bilinguals make what Weinreich (1953, p. know from observation, a small percentage of
7) has called “interlingual identifications.” In learners –perhaps a mere 5 percent. It follows
his book Languages in Contact, Weinreich from this assumption that this 5 percent go
discusses –though briefly- the practical need through very different psycholinguistic
for assuming in studies of bilingualism as that processes than do most second language
of a phoneme in two languages, or that of a learners and that these successful learners
grammatical relationship in two languages, or may be safely ignored –in a counterfactual
that of a semantic feature in two languages, sense 6- for the purposes of establishing the
have been made by the individual in question constructs which point to the psychologically
in a language contact situation. Although relevant data pertinent to most second
Weinreich takes up many linguistic and some language learners (i.e., the vast majority of
psychological questions, he leaves completely second language learners who fail to achieve
open questions regarding the psychological native speaker competence), the notion of
structure within which we assume “interlingual attempted learning is independent of and
identifications” exist; we assume that there is logically prior to the notion of successful
such a psychological structure and that it is learning. In this paper, we will focus on
latent in the brain, activated when one attempted learning by this group of learners,
attempts to learn a second language. successful or not, and will assume that they
activate a different, though still genetically
The closest thing in the literature to the determined structure (referred to here as the
concept latent psychological structure is the latent psychological structure) whenever they
concept latent language structure (Lenneberg attempt to produce a sentence in the second
1967, especially pp. 347-79) which, according language, that is whenever they attempt to
to Lenneberg, (a) is an already formulated express meanings, which they may already
arrangement in the brain (b) is the biological have, in a language which they are in the
counterpart to universal grammar, and (c) is process of learning.
transformed by the infant into the realized
structure of a particular grammar in This series of assumptions must be made, I
accordance with certain maturational stages. think, because the second language learner
For the purposes of this article, I will assume who actually achieves native speaker
the existence of the latent language competence cannot possibly have been taught
structured described by Lenneberg: I shall this competence, since linguistics are daily –in
further assume that there exists in the brain almost every generative study- discovering
an already formulated arrangement which for new and fundamental facts about particular
most people is different from and exists in languages. Successful learners, in order to
addition to Lenneberg’s latent language
structure. It is important to state that with the
latent structure described in this paper as 5
First pointed out by Harold Edwards.
6
compared to Lenneberg’s, there is no genetic See Lawler and Selinker (1979) where the relevance
of counterfactuals to a theory of second language
learning is taken up.
45
achieve this native speaker competence, must that TL. These three sets of utterances or
have acquired these facts (and most probably behavioral events are, then, in this
important principles of language organization) framework, the psychological relevant data of
without having explicitly been taught them. second language learning, and theoretical
predictions in a relevant psychology of second
Regarding the ideal second language learner language learning will be the surface
who will not succeed (in the absolute sense structures of IL sentences.
described above) and who is thus
representative of the vast majority of second By setting up these three sets of utterances
language learners, we can idealize that from within one theoretical framework, and by
the beginning of his study of a second gathering as data utterances related to
language, he has his attention focused upon specific linguistic structures in each of these
the norm of the language whose sentences he three systems, (under the same experimental
is attempting to produce. With this statement, conditions, if possible) the investigator in the
we have idealized the picture we wish to psychology of second language learning can
sketch in the following ways: the generally begin to study the psycholinguistic processes
accepted notion target language (TL), i.e., the which establish the knowledge which underlies
second language the learner is attempting to IL behavior. I would like to suggest that there
learn, is here restricted to mean that there is are five central processes (and perhaps some
only one norm of one dialect within the additional minor ones), and that they exist in
interlingual focus of attention of the learner. the latent psychological structure referred to
Furthermore, we focus our analytical attention above. I consider the following to be
upon the only observable data to which we processes central to second language
can relate theoretical predictions: the learning: first, language transfer; second,
utterances which are produced when the transfer of training; third, strategies of second
learner attempts to say sentences of a TL. language learning; fourth, strategies of
This set of utterances for most learners of a second language communication; and fifth,
second language is not identical to the overgeneralization of TL linguistic material.
hypothesized corresponding set of utterances Each of the analyst’s predictions as to the
which would have been produced by a native shape of IL utterances should be associated
speaker of the TL had he attempted to with one or more of these, or other,
express the same meaning as the learner. processes.
Since we can observe that these two sets of
utterances are not identical, then in the Fossilization
making of constructs relevant to a theory of
second language learning, one would be Before briefly describing these psycholinguistic
completely justified by hypothesizing, perhaps processes, another notion I wish to introduce
even compelled to hypothesize, the existence for the reader’s consideration is the concept of
of a separate linguistic system based on the fossilization, a mechanism which is assumed
observable output which results from a also to exist in the latent psychological
learner’s attempted production of a TL norm. structure described above. Fossilizable
This linguistic system we will call linguistic phenomena are linguistic items,
interlanguage (IL). One of the main points of rules, and subsystems which speakers of a
this article is the assumption that predictions particular NL will tend to keep in their IL
of behavioral events in a theory of second relative to a particular TL, no matter what the
language learning should be primarily age of the learner or amount of explanation
concerned with the linguistic shapes of the and instruction he receives in the TL. I have in
utterances produced in ILs. Successful mind such fossilizable structures as the well-
predictions of such behavioral events in known “errors”: French uvular /r/ in their
meaningful performance situations will add English IL, American English retroflex /r/ in
credence to the theoretical constructs related their French IL, English rhythm in the IL
to the latent psychological structure discussed relative to Spanish, German time-place order
in this paper. It follows from the above that after the verb in the English IL of German
the only observable data from meaningful speakers, and so on. I also have in mind less
performance situations we can establish as well known “non-errors” such as Spanish
relevant to interlingual identifications are: (1) monophthong vowels in the IL of Spanish
utterances in the learner’s native language speakers relative to Hebrew, and Hebrew
(NL) produced by the learner; (2) IL object-time surface order after the verb in the
utterances produced by the learner; and (3) IL of Hebrew speakers relative to English.
TL utterances produced by native speakers of Finally, there are fossilizable structures that
46
are much harder to classify such as some transfer; if these fossilizable items, rules, and
features of the Thai tone system in the IL of subsystems are a result of identifiable items in
Thai speakers relative to English. It is training procedures, then we are dealing with
important to note that fossilizable structures the process known as the transfer of training;
tend to remain as potential performance, if they are a result of an identifiable approach
reemerging in the productive performance of by the learner to the material to be learned,
an IL even when seemingly eradicated. Many then we are dealing with strategies of second
of these phenomena reappear in IL language learning; if they are a result of an
performance when the learner’s attention is identifiable approach by the learner to
focused upon new and difficult intellectual communication with native speakers of the TL,
subject matter or when he is in a state of then we are dealing with strategies of second
anxiety or other excitement, and strangely language communication; and finally, if they
enough, when he is in a state of extreme are a result of a clear overgeneralization of TL
relaxation. Note that the claim is made here rules and semantic features, then we are
that, whatever the cause, the well-observed dealing with the overgeneralization of TL
phenomenon of backsliding by second linguistic material. I would like to hypothesize
language learners from a TL norm is not, as that these five processes which are central to
has been generally believed, either random or second language learning, and that each
toward the speaker’s NL, but toward an IL process forces fossilizable material upon
norm. surface IL utterances, controlling to a very
large extent the surface structures of these
A crucial fact, perhaps the most crucial fact, utterances.
which any adequate theory of second
language learning will have to explain is this Combinations of these processes produce
regular reappearance or reemergence in IL what we might term entirely fossilized IL
productive performance of linguistic structures competences. Coulter (1968) presents
which were thought to be eradicated. This convincing data to demonstrate not only
behavioral reappearance is what has led me to language transfer but also a strategy of
postulate the reality of fossilization and ILs. It communication common to many second
should be made clear that the reappearance language learners. This strategy of
of such behavior is not limited to the phonetic communication dictates to them, internally as
level. For example, some of the subtlest input it were, that they know enough of the TL in
information that a learner of a second order to communicate. And they stop
language has to master regards sub learning. Whether they stop learning entirely
categorization notions of verbal or go on to learn in a minor way, e.g., adding
complementation. Indian English as an IL with vocabulary as experience demands (Jain
regard to English seems to fossilize the that (1969) insists they must) is, it seems to me, a
complement or V that construction for all moot point. If these individuals do not also
verbs that take sentential complements. Even learn the syntactic information that goes with
when the correct form has been learned by lexical items, then adding a few new lexical
the Indian speaker of English, this type of items, say on space travel, is, I would argue,
knowledge is the first he seems to lose when of little consequence. The important thing to
his attention is diverted to new intellectual note with regard to the evidence presented in
subject matter or when he has not spoken the Coulter (1968) and Jain (1969) is that not
TL for even a short time. Under conditions only can entire IL competences be fossilized in
such as these, there is a regular reappearance individual learners performing in their own
of the that complement in IL performance for interlingual situation, but also in whole groups
all sentential complements. of individuals, resulting in the emergence of a
new dialect (here Indian English), where
Five central processes fossilized IL competences may be the normal
situation.
It is my contention that the most interesting
phenomena in IL performance are those We will now provide examples of these
items, rules, and subsystems which are processes. The examples presented in the last
fossilizable in terms of the five processes section are almost certainly the result of the
listed above. It can be experimentally process of language transfer. A few examples
demonstrated that fossilizable items, rules, relating to the other processes should suffice
and subsystems which occur in IL for this paper.
performance are a result of the NL, then we
are dealing with the process of language
47
Overgeneralization of TL rules is a contrastive analysis then there should be
phenomenon well-known to language no trouble. It seems to be the case that
teachers. Speakers of many languages could the resultant IL form, in the first instance,
produce a sentence of the following kind in is due directly to the transfer of training;
their English IL: textbooks and teachers in this interlingual
situation almost always present drills with
1. What did he intended to say? he and never with she. The extent of this
fossilization can be seen with respect to
Where the past tense morpheme –ed is speakers of this IL over the age pf
extended to an environment in which, to eighteen, who even though they are
the learner, it could logically apply, but consciously aware of the distinction and of
just does not. The Indian speaker of their recurrent error, in fact, regularly
English who produces the collocation drive produce he for both he and she, stating
a bicycle in his IL performance, as in 2: that they feel they do not need to make
this distinction in order to communicate.
2. After thinking little I decided to start on In this case, then, the fossilizable error is
the bicycle as slowly as I could as it was due originally to a type of transfer of
not possible to drive fast. training and later to a particular strategy
of second language communication.
Is most probably overgeneralizing the use
of drive to all vehicles (Jain 1969, pp. 22 Concerning the notion strategy little is
and 24). Most learners of English quickly known in psychology about what
learn the English rule of contraction which constitutes a strategy; and a viable
forms things like the concert’s from the definition of it does not seem possible at
concert is, but then these learners may present. Even less is known about
overgeneralize this rule to produce strategies which learners of a second
sentences like: language use in their attempt to master a
TL and express meanings in it. It has been
3. Max is happier than Sam’s these days. pointed out that learner strategies are
probably culture bound to some extent.
In their English IL. Though this sentence is For example, in many traditional cultures,
hypothetical, it illustrates an earlier point. chanting is used as a learning device,
The learner of English who produces clearly relating to what is learned in these
contractions correctly in all environments situations. Crucially, it has been argued
must have learned the following constraint that strategies for handling TL material
without explanation and instruction, since evolve whenever the learner realizes,
this constraint was discovered only either consciously or subconsciously, that
recently: “contraction of auxiliaries… he has no linguistic competence with
cannot occur when a constituent regard to some aspect of the TL. It cannot
immediately following the auxiliary to be be doubted that various internal strategies
contracted has been deleted,” e.g., happy on the part of the second language learner
in (3) (Lakoff, 1971). Dozens of examples affect to a large extent the surface
of overgeneralization of TL rules are structures of sentences underlying IL
provided in Richards (1970). utterances. But exactly what these
strategies might be and how they might
The transfer of training is a process which work is at present pure conjecture. Thus,
is quite different from language transfer one can only roughly attribute the source
(see Selinker 1969) and from of the examples presented herein to one
overgeneralization of TL rules. It underlies or another strategy.
the source of a difficulty which Serbo-
Croatian speakers at all levels of English One example of a strategy of a second
proficiency regularly have with the he/she language learning is that is widespread in
distinction, producing in their English IL he many interlingual situations is a tendency
on almost every occasion wherever he or on the part of learners to reduce the TL to
she would be called for according to any a simpler system. According to Jain (1969,
norm of English. There is no language pp. 3 and 4), the results of this strategy
transfer effect here since, with regard to are manifested at all levels of syntax in
animateness, the distinction between he the IL of Indian speakers of English. For
and she is the same in Serbo-Croatian as example, if the learner has adopted the
it is in English. According to a standard strategy that all verbs are either transitive
48
 or intransitive, he may produce IL forms they call probability matching, where the
such as: chance that the learner will select an
alternative morphological ending related to
4. I am feeling thirsty. the cue noun is not random. Crothers and
Suppes do not provide examples of the result
Or of this strategy in meaningful performance
situations; an example would be the r at the
5. Don’t worry, I’m hearing him. end of words like California and saw which
foreign students of English who have had
And in producing them seems to have teachers from the Boston area regularly
adopted the further strategy that the reproduce in their English IL.
realization of the category aspect in its
progressive form on the surface is always To conclude this section, it should be pointed
with –ing marking (for further discussion, out that beyond the five so-called central
see Jain 1969, p. 3ff.). processes; there exist many other processes
which account to some degree for the surface
Coulter (1968) reports systematic errors form of IL utterances. One might mention
occurring in the English IL performance of spelling pronunciations, e.g., speakers of
two elderly Russian speakers of English, many languages pronounce final –er on
due to another strategy which seems also English words as /ε/ plus some form of r;
to be widespread in many interlingual cognate pronunciation, e.g., English athlete
situations: a tendency on the part of pronounced as /atlit/ by many Frenchmen
second language learners to avoid whether or not they can produce /θ/ in other
grammatical formatives such as articles English words; holophrase learning (Jain
(6), plural forms (7), and past tense forms 1969), e.g., for half-an-hour the Indian
(8): learner of English may produce one and half-
an-hour; hypercorrection, e.g., the Israeli who
6. It was Ø nice, nice trailer, Ø big one. in attempting to get rid of his uvular fricative
(Coulter 1968, p. 22) for English retroflex /r/ produces /w/ before
front vowels, “a vocalization too far forward”,
7. I have many hundred carpenter my own. and most assuredly others such as long
(Coulter 1968, p. 29) exposure to signs and headlines which
according to Jain (1969) affect by themselves
8. I was in Frankfort when I fill application. the shape of English IL utterances of Indians,
(Coulter 1968, p. 36). or at least reinforce more important processes
such as language transfer.
This tendency could be the result of a learning
strategy of simplification, but Coulter (1968, Problems with This Perspective
p. 7 ff.) attributes it to a communication
strategy due to the past experience of the There are certainly many questions one might
speaker which has shown him that if he thinks wish to ask regarding the perspective
about grammatical processes while attempting presented so far in this article; I shall attempt
to express in English meanings which he to deal with five. The reader should bear in
already has, then his speech will be hesitant mind that we are here calling for the
and disconnected, leading native speakers to discovery, description and experimental
be impatient with him. Also, Coulter claims testing of fossilizable items, rules and
that this strategy of second language subsystems in interlanguages and the relating
communication seemed to dictate to these of these to the above-mentioned processes –
speakers that a form such as the English especially to the central ones. What seems to
plural “was not necessary for the kind of be most promising for study
communicating they used” (Coulter 1968, p.
30). Is the observation concerning fossilization.
Many IL linguistic structures are never really
Not all of these strategies, it must be pointed eradicated for most second language learners;
out, are conscious. A subconscious strategy of manifestations of these structures regularly
second language learning called cue-copying reappear in IL productive performance,
has been experimented with by Crothers and especially under conditions of anxiety, shifting
Suppes (1867, p. 211) on Americans learning attention, and second language performance
Russian morphological concepts. This copy the on subject matter which is new to the learner.
cue strategy is most probably due to what In this observation which allows us to claim
49
that these psycholinguistic structures, even “How do I recognize fossilizable structures in
when seemingly eradicated, are still somehow advance?” or “Why do some things fossilize
present in the brain, stored by a fossilization and others do not?,” all experiments
mechanism (primarily through one of these conducted within the framework provided in
five processes) in an IL. We further this paper must be regarded as exploratory in
hypothesize that interlingual identifications, nature. (To put things in more familiar jargon:
uniting the three linguistic systems (NL, IL, with regard to fossilization, our results are
and TL) psychologically, are activated in a descriptive and not explanatory in nature.)
latent psychological structure whenever an But this task of prediction may prove to be
individual attempts to produce TL sentences. impossible; certainly as Fred Lukoff point out
(personal communication) this task, on the
The first problem we wish to deal with is: can face of it, may be even tougher than trying to
we always unambiguously identify which of predict errors in second language performance
these processes our observable data is to be –a task notably lacking in success.
attributable to? Most probably not. It has been
frequently pointed out that this situation is The major justification one has for writing
quite common in psychology. In studies on about the construct fossilization at this stage
memory, for example, one often does not of knowledge is that descriptive knowledge
know whether a particular constituent IL about ILs which turns out to suggest
concatenation is a result of language transfer predictions verifiable in meaningful
or of transfer of training or, perhaps, of both. performance situations, leads the way to a
But this limitation need not deter us, even if systematic collection of the relevant data; this
we cannot always sort things out absolutely. task, one which is impossible without this
By applying the constructs suggested in this construct, is expected to be relevant to
paper, I believe that relevant data can be serious theory construction in a psychology of
found in the very many second language second language learning.
learning situations around us.
The third problem to be treated here concerns
The second problem is: how can we the apparent difficulty of fitting the following
systematize the notion fossilization so that type of question into the idealized domain I
from the basis of theoretical constructs, we have been sketching: how does a second
can predict which items in which interlingual language learning novice become able to
situations will be fossilized? To illustrate the produce IL utterances whose surface
difficulty of attempting to answer this constituents are correct, i.e., correct with
question, note in the following example the respect to the TL whose norm he is
nonreversibility of fossilization effects for no attempting to produce? This question finally
apparent reason. According to a contrastive brings us face to face with the notion of
analysis, Spanish speakers should have no success in absolute terms: productive
difficulty with the he / she distinction in performance in the TL by the second language
English, nor should English speakers have any learner which is identical to that produced by
difficulty with the corresponding distinction in the native speaker of that TL. We noted this
Spanish. The facts are quite different, earlier so as to exclude from or idealized
however: Spanish speakers do, indeed, domain of inquiry those learners of second
regularly have trouble with this distinction, languages who reactivate the latent language
while the reverse does not seem to occur with structure that is realized into a native
English learners of Spanish. Unlike the Serbo- language. In this article, we are concentrating
Croatian example mentioned above, in this on attempted learning of a second language,
case there is no clear-cut explanation why unsuccessful in this absolute sense. Of course,
Spanish speakers have trouble and English success in second language learning need not
speakers do not. In cases such as these, it to be defined so absolutely. The teacher or
may turn out that one process, e.g., language the learner can be satisfied with the learner’s
transfer or transfer of training, overrides other achieving what has been called
considerations, but the stating of the “communicative competence” (see, for
governing conditions may prove very difficult example, Jakovits 1970, or Hymes 1972). But
indeed. this is not the issue here. As was pointed out
in the introduction, the emphasis upon what
In principle, one feels forced to agree with the teacher has to do in order to help the
Stephanie Harries (personal communication) learner achieve successful learning belongs to
who claims that until a theory of second the teaching perspective, which is not the
language learning can answer questions like: perspective of this paper. Perhaps the rather
50
curious confusion in the literature of learning Concerning underlying linguistic structure, we
a second language with teaching a second should perhaps not be too surprised of it turns
language can be explained by the failure to out not to matter whose model we need, if an
see a psychology of second language learning eclectic one will do, or even if such notions as
in terms other than those related to success. the cycle, tree pruning, or even derivation
For example, typical learning theory prove not to have much relevance. If it is
experiments when done in the domain of reasonable to assume that the only
second language learning would demand linguistically relevant unit of a theory of
knowledge or where the learner would like second language learning is one which is
him to end up. Experiments of this type would identified interlingually across three linguistic
also demand knowledge of where the second systems (NL, TL, and IL) by means of
language learner begins. We would claim that fossilization and the processes described
prerequisite to both these types of available to earlier, then it follows that no unit of linguistic
us. Thus, such experiments at present are theory, as these units are currently conceived,
premature, with the results bound to prove could fit this criterion. More generally, we
confusing. should state that there is no necessary
connection between relevant units of linguistic
Specifically concerning the problem raised in theory and linguistically relevant units of a
the first sentence of this section, it seems to psychology of second language learning. That
me that this question, though relevant to the this assumption is obviously correct is clear to
psychology of second language learning, is me; that many linguists will not be convinced
one that should also not be asked for the is also clear.
present since its asking depends upon our
understanding clearly the psychological extent For evidence of the relevant unit of surface
of interlingual identifications. For example, syntactic structure, applying at one and the
before we can discover how surface same time to these three linguistic systems, I
constituents in an IL get reorganized to refer the reader to experimental evidence
identify with the TL, we must have a clear appearing in my paper on language transfer
idea of what is in that IL, even if we cannot (Selinker 1969). In those experiments
explain why it is there. In Selinker (1969) I subjects responded orally in their native
believe I have shown that within a very language to questions presented orally in their
limited interlingual situation, the basis from NL and attempted to respond in English to
which linguistic material must be reorganized parallel questions presented in English. The
in order to be correct has been operationally questions came from an interview designed to
and unambiguously established. But I have elicit manifestations of specific types of
there said nothing about the way in which surface structures in certain syntactic
successful learners do in fact reorganize domains. The only experimental instruction
linguistic material from this particular IL. Here given was for each subject to speak in a
we can speculate that as part of a definition of “complete sentence.” Replicated results
learning a second language for most learners, showed that the interlingual unit of surface
involves, to a large extent, the reorganization syntactic structure transferred from NL to IL
of linguistic material from an IL to identity (not to TL) was a unit roughly equivalent to
with a particular TL. the traditional direct object or to an adverb of
place, and adverb of degree, and so on. I
The fourth problem is: (a) what are the would claim that this unit, a surface
relevant units of this hypothezised latent constituent labeled the syntactic string, has a
psychological structure within which behavioral unity both in the experimental
interlingual identifications exist and (b) is situation and in meaningful performance
there any evidence for the existence of this situations, and thus, if the results were
units? If the relevant data of the psychology replicated in other interlingual situations (i.e.,
of second language learning are in fact other combinations of NL, TL, and IL), would
parallel utterances in three linguistic systems account for a large class of IL events.
(NL, IL, and TL), then it seems to me
reasonable to hypothesize that the only With regard to a realizational unit, i.e., a
relevant, one might say, psychologically real, syntactic string tied to a specific semantic
interlingual unit is one which can be described notion, replicated results from this same
simultaneously for parallel data in the three series of experiments show that responses
systems, and, if possible, for experimentally concerning a topic such as “subjects studied in
induced data in those systems. school,” as opposed to other topics such as
“buying and receiving things” and “seeing
51
movies and parades,” affected very drastically “the unit” in native-speaker speech perception
the surface concatenation of the above- is a waste of time. Alternative units may be
mentioned strings. This semantic effect on available to native speakers, for example
surface syntactic order in an interlingual under noise conditions. While other
study, if further replicated in other interlingual explanations are surely possible for the well-
situations, would provide very powerful known fact that noise conditions affect
evidence for the transfer of the whole performance in a second language, and
realizational unit as well as for its candidacy sometimes drastically, we cannot ignore the
as the unit of realizational structure in possible relevance of Haggard’s intriguing
interlingual identifications. suggestion: that alternative language units
are available to individuals and that these
Concerning the notion of relevant units on the units are activated under certain conditions. It
phonological level, it seems to me that Briere fits in very well with the perspective outlined
(1968) has demonstrated that for this data in this paper to postulate a new type of
there are several relevant units. The relevant psycholinguistic unit, available to an individual
units do not always correspond to known whenever he attempts to produce sentences
linguistic units, but rather would depend on in a second language. This interlingual unit
the sounds involved; sometimes the stretches, we hypothesize, across three
taxonomic phoneme is the unit, but the unit in linguistic systems: NL, IL, and TL, and
other cases seems not to be describable in becomes available to the idealized second
purely linguistic terms. Briere evolved an language learner who will not achieve native-
experimental technique which imitated to a speaker competence in the TL, whenever he
large extent actual methods of teaching attempts to express meanings, which he may
advocated by applied structural linguists: already have, in a TL he is learning, i.e.,
listening to TL sounds, attempted imitation, whenever he attempts to produce a TL norm.
use of phonemic transcription, physiological these units become available to the learner
explanations, and so on. If I may be allowed only after he has switched his psychic set or
to reinterpret Briere’s data, it seems to me state from the native-speaker domain to the
that he has been working, in another new domain of interlingual identifications. I
interlingual situation, with exactly the three would like to postulate further that these
systems we are discussing here, NL, TL, and relevant units of interlingual identifications do
IL: first, NL utterances which were not come from anywhere; they are latent in
hypothesized utterances in American English; the brain in a latent psychological structure,
second, TL utterances which were actual available to an individual whenever he wishes
utterances in the composite language Briere to attempt to produce the norm of any TL.
set up, each utterance having been produced
by a native speaker of French, Arabic, or The final difficulty with this perspective which
Vietnamese; third, IL utterances which were we will treat here is the following: how can we
actual utterances produced by native speakers experiment with three linguistic systems,
of this NL when attempting to produce this creating the same experimental conditions for
particular TL norm. regarding the sounds /ž/ each, with one unit which is identified
and /ŋ/ in this TL corpus, the unit identified interlingually across these systems? I can only
interlingually across these three systems is refer the reader once again to my own
the taxonomic phoneme defined experiments on language transfer (Selinker
distributionally within the syllable as opposed 1969) where manifestations of desired
to within the word (Briere 1968, p. 73). For concatenations of particular surface syntactic
other sounds the relevant phonological unit of structures were obtained in what, I believe,
interlingual identifications is not the was an efficient and valid manner. An oral
taxonomic phoneme, but may be based on interview technique was used; the purpose of
phonetic parameters some of which, he says, the interview was to achieve a similar
are probably not known (Briere 1968, pp. 73 framework in the three systems which served
and 64). the interviewer as a guide in his attempt to
elicit certain types of sentences from the
If these units in the domain of interlingual subjects. Upon request, I am prepared to
identifications are not necessarily the same make available a transcript of this interview as
units as those in the native-speaker domain, well as some thoughts for its improvement.
then where do they come from? An interesting Future experimental work, to be undertaken
bit of speculation about native-speaker within this perspective, will go toward
performance units is provided by Haggard investigating the kind and extent of linguistic
(1967, p. 335) who states that searching for
52
structures amenable to this particular 12. The syntactic string is the unit of surface
technique. structure transfer and part of the unit of
realizational transfer.
Summary 13. The taxonomic phoneme is, in the case of
some sounds, the unit of interlingual
The following are some assumptions which are phonology, while in other cases no purely
necessary for research into the linguistic linguistic unit seems relevant.
aspects of the psychology of second language 14. there exists a latent psychological
learning and which have been suggested by structure, i.e., an already formulated
the above discussion. arrangement in the brain, which is
activated wherever an adult attempts to
1. In a theory of second language learning, produce meanings, which he may have, in
those behavioral events which are to be a second language which he is learning.
counted as relevant data are not 15. Interlingual identifications, the units
immediately obvious. mentioned in 12 and 13, and the
2. These data have to be organized with the processes listed in 8 exist in this latent
help of certain theoretical constructs. psychological structure.
3. Some theoretical constructs relevant to 16. Fossilization, a mechanism which also
the way in which adults actually learn exists in this latent psychological
second languages are: interlingual structure, underlies surface linguistic
identifications, native language (NL), material which speakers will tend to keep
target language (TL), interlanguage (IL), in their IL productive performance, no
fossilization, syntactic string, taxonomic matter what the age of the learner or the
phoneme, phonetic feature. amount of instruction he receives in the
4. The psychologically relevant data of TL.
second language learning are utterances 17. The fossilization mechanism accounts for
in TL by native speaker, and in NL and IL the phenomenon of the regular
by second language learners. reappearance in IL productive
5. Interlingual identification by second performance of linguistic material which
language learners is what unites three was thought to be eradicated.
linguistic systems (NL, TL, and IL) 18. This latent psychological structure, for
psychologically. These learners focus upon most learners, is different from and exists
one norm of the TL. in addition to the latent language structure
6. Theoretical predictions in a relevant described by Lenneberg (1967, pp. 374-
psychology of a second language must be 79).
the surface structures of IL sentences. 19. These two latent structures differ in the
7. Successful second language learning, for following ways: (a) the latent
most learners, is the reorganization of psychological structure has no genetic
linguistic material from an IL to identity timetable; (b) it has no direct counterpart
with a particular TL. to any grammatical concept; (c) it may
8. There exist five distinct processes which not be activated at all; (d) it may never be
are central to second language learning: realized into a natural language; and (e) it
language transfer, transfer of training, may overlap with other intellectual
strategies of second language learning, structures.
strategies of second language 20. The qualification (“for most learners”) in 7
communication, and overgeneralization of and 18 is necessary, since those adults
TL linguistic material. who seem to achieve native-speaker
9. Each prediction in 6 should be made, if competence, i.e., those who learn a
possible, relative to one of the five second language so that their performance
processes in 8. is indistinguishable from that of native
10. There is no necessary connection between speakers (perhaps a mere 5 percent of all
relevant units of linguistic theory and learners), have not been taught this
linguistically relevant units of a psychology performance through explanation and
of second language learning. instruction but have somehow reactivated
11. The only linguistically relevant unit of a this latent language structure.
psychology of second language learning is 21. Since it is assumed that the two structures
one which is identified interlingually across mentioned in 18 are different and since we
the three linguistic systems: NL, TL, and know very little about the latent language
IL. structure and its activation, then the 5
percent mentioned in 20 should be ignored
53
 in setting up the idealizations which guide
us to the psychologically relevant data of such as go in, go out, go up, go away, go
second language learning. over, etc, as in conversation-level equivalents
of enter, leave, ascend, depart, survey or
Readings in SPANISH-ENGLISH cross, etc. The verbs DO and MAKE are active
Contrastive Linguistics in these derivations and, as might be
expected, the meanings in Spanish range over
Edited by Rose Nash a wide semantic area.

Make and do in phrasal verbs In these phrasal verbs the particle often
carries more information than the verb (in go
Explanation before doing exercises: Many up and go down it is up and down that really
English verbs (including some of the most carry the message), and this is especially true
common and most frequently used ones) are with DO, which is the prototypical pro-verb.
made up of two (or more) words. These are Also one notes that some particles will appear
not always found in dictionaries, since with only one of the pair DO and MAKE, other
dictionaries specialize in single-word entries. with either. Study the following presentation
But informal English is replete with examples, in preparation for a vocabulary exercise:

Phrasal verb Paraphrase Translation Example

With DO only
Do in Kill Matar They’ll do him in if they get a chance.
After the long race he was really done
Do(ne) in Exhausted Agotado
in.
Do with Deal with Ver con I’ll have nothing to do with him.
Do out of Cheat Defraudar He did me out of ten dollars.
Do away with Get rid of Suprimir He did away with the evidence.
Managed
Do without Manejarse sin He did without food for days.
not having
With MAKE only
Succeed Salir How did you make out?
Make out
Understand Interpretar He couldn’t make out the answer.
Make off Escape Largarse He made off into the night.
Make off with Take away Llevarse He made off with all her money.
Make after Pursue Perseguir He made after the thief.
Think,
Make of Creer, entender What do you make of it?
understand
Make as if Pretend fingir Make as if you don see him.
Make up for Compensate Compensar He made up for his losses.
Be friendly
Make up to Congraciarse con He made up to the secretary.
toward
With either DO or MAKE
Robert made up some excuse.
Invent Inventar
Ann made up a batch of cookies
Bake Cocer al horno
Make up John was absent but he made up his
Catch up Recuperar
assignment.
Straighten Arreglar
Make up my room early, please.
Close Cerrar Will you do up my zipper.
Tie Amarrar Please do this package up.
Do up
Arrange Arreglar He did everything up fine.
Complete Terminar That does it up just fine.
Make over Redesign Modificar Mother made dress over for the child.
Do over Redecorate Renovar She wants to do the house over.
Head
Make for Dirigirse hacia He made for the kitchen.
towards
Ruined,
Do(ne) for Arruinado He’s all done for.
finished

54
It should be noted that stress patterns 21. I don’t know what I’d __________ without
involved in phrasal verbs should be taught, or you.
at least correctly modeled if similar stress 22. I suppose somehow I’ll __________ out.
patterns have been presented previously. 23. But since somebody __________ off with
There is more than one pattern, as can be my car, I’m lost.
seen by comparing make off with make of. 24. I’ve been __________ out of things
before, but never anything so big!
Exercise 25. I don’t know how I’ll __________ up for
my foolishness in leaving the key in the
Instructions: examine each sentence, decide ignition.
whether DO or MAKE is the appropriate word
for the blank, and then read the completed Exercise
sentence. You are learning the correct phrasal
verb for each situation, but in this exercise Instructions: Examine each sentence, decide
the particle after the blank can serve as an whether DO or MAKE appropriately fits in the
adequate cue. Since you are selecting a verb, blank, then read the completed sentence.
be sure to pick an appropriate tense form. Unlike the preceding exercise, the particles in
Note that the particle may be separated from the present sentences will not serve as cues,
the verb. since these particles can go with either DO or
MAKE. You will have to make your selections
1. What does the governor have to consistent with the larger context of the entire
__________ with a court decision? sentence. Note that the particle may be
2. He was __________ in by the mob. separated from the verb.
3. The prisoner slipped through the broken
window and __________ off. 1. His excuses all sound __________ up.
4. He __________ off with the chairman’s 2. Could you please __________ up my
car. zipper?
5. He __________ the widow out of her life’s 3. I’ve just got time to __________ up a
savings. batch of cookies.
6. __________ as if you didn’t hear the 4. Have you __________ up the assignment
confession. you missed?
7. You absolutely must __________ away 5. Hang this sign on your door if you want
with this letter. the maid to __________ up your room.
8. I couldn’t __________ out what she said 6. You’ll have to __________ up that
over the phone. package better or the post office won’t
9. He ran at least ten miles; that would accept it.
__________ anybody in. 7. This dress is too old; it’s not worth trying
10. The kidnapper __________off with the to __________ it over.
baby. 8. I’ve got to find somebody to __________
11. The gamble paid off; he __________ up this room over.
for his losses. 9. He’s __________ for; he’ll never fight
12. How did you __________ out on your professionally again.
safari? 10. The thief __________ right for the safe.
13. How he __________ without food for ten 11. Mary __________ up a batch of chocolate
days is a mystery.. fudge.
14. He’s always trying to __________ up to 12. She spilled the salt in it and had to
the teacher. __________ it over.
15. How did you __________ out in the 13. It seems like everything she does, she has
poetry competition? to __________ over.
16. It’s a mystery to me; what do you 14. She should have forgotten the kitchen and
__________ of his answer? have tried to __________ up her late
17. I’ll never again have anything to school assignment.
__________ with the stock market. 15. She’ll be __________ for if she doesn’t
18. The barking hounds __________ after the take her class work more seriously.
fox. 16. But she likes to cook, so instead of
19. Don’t let some lawyer __________ you studying she just __________ for the
out of your inheritance. kitchen.
20. I don’t know what to __________ of her 17. I suppose it’s unless to try to __________
answer. her over.

55
18. She’d rather __________ up cookies than White phone Bone white
__________ up homework. Spanish examples
19. Though she’s also good at __________ up Pobre niño (unfortunate Niño pobre (poor boy)
boy)
excuses. Alta dama (important lady) Dama alta (tall lady)
20. That she __________ up brown. Gran ciudad (great city) Ciudad grande (large
city)
Word Order: A Problematic Aspect of c) Word order changes the meaning of
Teaching English as a Second Language certain modifiers.
When words such as good, pretty,
Zenobia Vélez Molina awful, are shifted from one position to
another, they change their meaning:
Eric H. Kadler says in his book, Linguistics and English adjectives English adverbs
Teaching Foreign Languages, Native speakers The plan is good The plan is good enough
may find it expedient to make Miss Perez is a pretty Miss Perez is a pretty
morphophonemic deviations from the woman smart woman
This is an awful bracelet This is an awfully
standard rules or they may make occasional expensive bracelet
semantic mistakes, but they seldom make
This phenomenon does not occur in
syntactic errors (syntactic encoding) to
Spanish since modifiers are marked by
produce ungrammatical sentences. Non-native
inflectional endings for adjectives and
speakers do make syntactic mistakes as
adverbs.
frequently and stubbornly as they make
II. Ungrammatical sequences.
semantic and morphological mistakes,
a) Reversing the position of nouns and
because they tend to transfer to the foreign
adjectives in English produces an
language their native syntactical system as
ungrammatical sequence.
well as their morphological habits and
Brown coat coat brown
semantic values.
Black dog dog brown
Big clock clock big
When a native speaker of English hears the
This condition does not produce
statement He eats everyday, he is unlikely to
ungrammaticality in Spanish. It only
conceive any other order or position for the
produces change of meaning.
worlds he, eats, everyday. If the statement
Carrying over word order the original
were changed to eats he everyday or
language may produce two possible
everyday eats he it would seem absurd.
results: first, it may change the
meaning to another one, unintended:
Word order change produces two possible
Example
results: it may change the meaning to another They also have a ladies’ También tienen un
meaning (both grammatical); or, it may department departamento de señoras
produce ungrammaticality. Also have a department of
I. Change of meaning. ladies
a) word order, together with intonation, Or it may produce an ungrammatical
distinguishes statements from sequence
questions: I have a red dress Tengo un vestido rojo
English statements English questions I have a dress red
She is a nurse. Is she a nurse? Robert Lado contrasts the position of
They are brothers. Are they brothers? single word modifiers in English and
He is in the park. Is he in the park? Spanish when he says:
Spanish statements Spanish questions In the structure of the noun phrase in
Ella es una bailarina. ¿Es ella bailarina? English, single word modifiers usually
Ellos son amigos. ¿son ellos amigos?
precede the head, and phrase and
El esta en la escuela ¿está él en la escuela?
clause modifiers follow it. In Spanish,
b) word order changes the meaning of
a normal position for all such modifiers
certain expressions:
English examples
is after the head. Compare the
Milk chocolate Chocolate milk Spanish casa “azul” with the English
Dog house House dog “blue” house.
Lamp oil Oil lamp Interference will commonly affect
Bus station Station bus Spanish-language students’ learning
Candle light Light candle process of English word order. The
Blue baby Baby blue contrastive position of single word
Red wine Wine red
modifiers in English and Spanish make
Pink rose Rose pink
Green Nile Nile green
this a difficult area, since the students

56
 tend to bring the linguistic habits of Because the student transfers the habits of his
their vernacular into the second … system there will be a problem when the
language. sound systems of English and Spanish differ.
Examples: Sometimes the difference will be phonemic,
The baby has blue eyes El bebé tiene ojos azules with the result that the student may say a
The baby has eyes blue word he does not intend to say, or he may
Luis is a tall boy Luis es un niño alto
hear a word that was not spoken to him.
Luis is a boy tall
Puerto Rico is a beautiful Puerto Rico es una Isla
Other times the difference will be sub-
island preciosa phonemic, so that the distortion results in a
Puerto Rico is an island foreign effect but not in a different word.
beautiful.
The concern with sound and sound contrasts
Reflexivization in Spanish motivates much of the phonetic drill material
found in our English as a second language
Mercedes Roldán textbooks and laboratory tapes, with the
objective of training student to avoid
The phenomenon of reflexivilization in English transferring speaking habits that may cause
has been studied at considerable length. The word confusions. Thus, a great deal of
first extensive analysis of the problem was attention is paid to such matters as the
done by Lees and Klima. They proposed a phonemic distinction between high front tense
condition for English reflexivilization: that a vowel /iy/ and high front lax vowel /I/, so that
noun-phrase which is coreferential with a word like sheep will not sound like ship; to
another noun-phrase mentioned earlier in the the proper lengthening of vowels before
same simplex sentence must be transformed voiced consonants so that they will not sound
into a reflexive pronoun. At the time, Lees and like voiceless consonants, as in pig-pick, and
Klima were not able to motivate their other pairs containing sounds that are
proposed rule fully, and many cases of it had allophonic in the student’s native language, so
to be accepted on faith. In the ten years that that, e.g., going to Yale will not be confused
have elapsed since then, as linguists have with going to jail.
gained a better understanding of deep
structure, most of those apparent exceptions This might be called the “Sound-to-Sound”
have been explained away and Lees and approach to phonological interference. In this
Klima’s analysis is one of the few early paper I shall discuss another kind of
transformational rules of English that has pronunciation error, one which results not
endured. from differences in the phonological systems
in two languages, but from differences in their
Orthographic Interference in orthographic systems that affect
Pronunciation pronunciation. Orthographic interference in
pronunciation occurs in the speech of students
Rose Nash who have learned to read and write their own
language before beginning the study of
Conventional studies of phonological English, and whose relative exposure to
interference, and teaching materials based on spoken English and written English has been
such studies, have dealt primarily with only highly disparate, rather than integrated.
one kind of pronunciation error –that which
stems from differences in the sound systems In one of my ESL classes, I have two
of the two languages involved. The classic intelligent students from the same part of
statement of the problem was made by Lado: South America. Student A never studied
English before coming to Puerto Rico a few
We know now that a speaker of one language years ago. However, being blessed with a
tends to transfer the entire system of his good ear and a retentive memory, he has
language to the foreign language:… he tends picked up enough English in a predominantly
to transfer his sound system, including the Spanish-speaking environment to use it daily
phonemes, the positional variants of the in his job as a tourist guide. His speech is by
phonemes. And the restrictions on no means accent-free, but it is quite
distribution. He tends to transfer his syllable intelligible, and we have no trouble
patterns, his word patterns, and his intonation communicating. His written compositions,
patterns as well. however, bear little resemblance to the
English that would be acceptable to Miss
Fidditch. But in fact, although they are full of
57
misspellings, they are rather good as patterns, or as spelling patterns. His visual
phonemic transcriptions. I read his memory for language, however, is still a
compositions aloud when I am correcting reflection of his more highly developed
them so I can recognize the words he auditory memory. He is, in effect, finding out
intended. For this student, English obviously what his spoken vocabulary looks like.
means spoken English. When he writes, the
sounds determine the spelling. In Stage Three, when the child can read and
write on his own, his visual memory starts to
Student B, on the other hand, studied English function independently of his auditory
for several years before coming to Puerto memory. He can now recognize words by their
Rico, using a grammar-translation method shapes alone, without having heard them
with heavy emphasis on reading previously.. and he can predict their
comprehension. His compositions would pronunciations, being guided by already
delight Miss Fidditch. His spellings are established associations between sound
flawless, and his vocabulary extensive. patterns and spelling patterns. In a language
However, when he speaks, it sounds like like Spanish, where there is a fairly close fit
Spanish gibberish, and I have great difficulty between phonology and spelling, the process
following him. He uses so many of acquiring literacy is completed at an early
pronunciations based on Spanish phoneme- age. In a language like English, however,
grapheme correspondences that many words Stage Three may continue into adulthood as
are distorted beyond recognition. For this associations made on the basis of past
student, English obviously means written experience continue to be revised for some
English; in speaking the spellings determine words. I can still remember my great surprise
the sounds. upon learning, in my second year of high
school, that the English word spelled s-u-c-c-
While these are probably extreme cases, I u-m-b was not pronounced /s«ks«m/ by
think they demonstrate that cross-language analogy with succeed. And it was not until I
phonological performance is definitely visited England at the ripe old age of 35 that I
influenced by orthography. This extends even realized, much to my chagrin, that Leicester
to the names of letters in the alphabet. The /laysestr/ and Leicester /lest«rare one and the
names of vowel letters are particularly same city.
confusing for my students, because English e
/iy/ is Spanish e /e/, English a /ey/ is Spanish The importance of Stage Three, that is, the
a /a/, and English I /ay/ is Spanish I /i/. Once, development of an independent visual
during a writing exercise in class, student A, memory for language, has not been properly
who was at the blackboard, misspelled the investigated. Yet it is clear that we use written
word please as plis. When I told him the language constantly in our daily lives, starting
correct spelling was p-l-e-a-s-e, he then wrote with the morning newspaper, and that when
pliesi. At this point student B volunteered to we want to remember something, we write it
set him straight, telling him to write what down. It may even be that the lexicon of a
sounded to me like “play ah say” and out language is stored in the mind on
came the desired spelling pattern. orthographic rather than phonemic form. If
you doubt that, try thinking of a place-name
In order to understand better how without remembering the spelling.
orthographic interference works in a cross-
language situation, it may be useful to review Now let us see what happens in a language-
the stages that a monolingual passes through learning situation, assuming that instruction in
in acquiring literacy. Stage One is exemplified English begins in Stage Three of native
by a typical pre-school child. In this stage, the Spanish acquisition.
child has only an auditory memory for
language. That is, he perceives, registers, and Given the command “Repeat X,” a beginning
recalls words solely in terms of sound images, student will produce Sound-to-Sound
which gradually become recurring patterns as interference. Thus, he may repeat the word
his vocabulary increases. school /sku:l/ as /eskul/, changing the initial
cluster, the vowel length and quality, and the
When the child begins to learn to read and allophonic character of the final consonant to
write his language, he enters Stage Two. He conform to Spanish articulatory patterns. In
then acquires visual images for the words he other words, he gives a Stage One response
already knows and can now perceive, register, that is inappropriate for English.
and recall them in two ways: as sound
58
Given the command !Write X,” the student will Mary is taller than Betty (like taller –
produce Sound-to-Spelling interference. Thus, workshop-)
he may write the word school /skuwl/ as skul /təlƏr/
or scul, representing the sounds he hears with /tayer/
the corresponding letters. In other words, he
gives a Stage Two response that is Second, there are identical graphemes in the
inappropriate for English. two languages that sometimes represent the
same phonemes. For example, ch in Spanish
Sound-to-Sound interference and Sound-to- chico matches ch in English cheese, but not in
Spelling interference start with an auditory English chemistry or machine. Some
image of the English word. If, however, the examples:
command is “Read X aloud,” the starting point
is not an auditory image, but a visual image, In general, television is pretty bad (like
that is, a spelling pattern from which he must general9
supply a pronunciation. In this case, he will /jenƏrƏl/
produce Spelling-to-Sound interference, and /heneral/
may say school /skuwl/ as /esčol/, reading ch My head aches (like hache ‘H’)
as /č/ and oo as /o/ as they wouild be in /eyks/
Spanish. In other words, he gives a Stage /eyčz/
Three response that is inappropriate in
English. Included in this group are all the vowel
graphemes. A great deal has been written
Of course no student, even a beginner, will lately about the regularities of English
ever make all three errors for the same word. spelling, but the fact remains that, vowel
And students do not speak or write English graphemes have highly variable values. Prator
only in response to commands. Therefore and Robinett in their Manual of American
teachers should be aware that when there is English Pronunciation list 12 different
an error in pronunciation, the source may be pronunciations for a, 17 for e, 8 for I, and 15
phonological interference, or orthographic for o when they appear alone or as the initial
interference, or a combination of the two. If a member of a vowel cluster. Spanish vowels
student says he is majoring in /čemistri/ (for have one pronunciation apiece. This almost
chemistry) you can be sure that he learned always matches one of the possible
this word in written form. pronunciations for the identical English vowel
grapheme. For example, o in Spanish lo
Pronunciation errors traceable to orthographic matches o in English go. The difficulty is that
interference fall into two broad and this one corresponding pronunciation is never
overlapping categories: those related to the most frequent one in English.
Spanish spelling patterns and those related to
inconsistencies in English spelling patterns – Third, there are vowel clusters pronounced as
inconsistencies, at least, from the point of separate syllables in Spanish but as single
view of the Spanish speaker. The words in vowels in English, such as ea in Spanish sea
English most susceptible are those with /sea/ and English sea /siy/. There are also
Spanish cognates, those with silent letters, some English vowel clusters, such as oo,
those resembling other words pronounced which are pronounced as a single syllable as
differently, homographs (i.e., same spelling, in coop /kuwp/ but as separate syllables
different pronunciation) and homophones across morpheme boundaries, as in cooperate
(i.e., same pronunciation, different spelling.) /koapəreyt/, but the student may treat both
the same, pronouncing each separately.
First, there are identical graphemes in the two
languages that never represent the same Fourth, there are silent letters. In Spanish
phonemes. These include j, ll, v, z, and qu. only h is silent, in English there is a wide
variety of letters that are sometimes silent,
Some examples: including h as in honest, but not in honey.
With liberty and justice for all (like justicia)
/jƏstis/ Fifth, there are word pairs identical in either
/hustis/ spelling or pronunciation, but not both, i.e.,
I quit my job (like quita) homographs and homophones. For example
/kwit/ bote and vote are both pronounced /bote/.
/kit/ But such pairs are infrequent, and never affect
vowels. The student who makes this kind of
59
error may have resigned himself to the fact
that some English words are not spelled the The Linguistic Components of Contrastive
way they should be spelled, but he assumes Analysis
that different spellings must have different
pronunciations. Homographic word pairs, on Contrastivists see it as their goal to explain
the other hand, will be pronounced identically. certain aspects of L2 learning. Their means
are descriptive accounts of the learner’s L1
The fifth type of error is found only among and the L2 to be learnt, and techniques for
advanced students –those who already have a the comparison of these descriptions. In other
good command of English phonology and words, the goal belongs to psychology while
orthography, so good, in fact, that they will the means are derived from linguistic science.
consciously avoid Spanish-sounding It is in fact this demarcation of goal and
pronunciations whenever an English spelling means, through their allocation to two
offers more than one possibility. different sciences, which disqualifies CA from
becoming subsumed under the rubric of the
What can we, as teachers, do about Spelling- hybrid discipline called ‘psycholinguistics’. I
to-Sound interference? First of all, we must shall argue later that some of the
remember that, to the student, his incorrect misunderstanding surrounding CA has arisen
pronunciations are the logical ones. He has from the mistaken view that CA is a form of
jumped into a foreign language at Stage psycholinguistics.
Three without the years of preparation in
Stages One and Two which were available to With certain notable exceptions (Firth, 1951)
him in his native language. We should give modern 20th century linguistics has seen as its
him credit for making educated guesses, even goal the description of the linguistic code,
though they turn out to be wrong. without making reference to the uses to which
the code is put, or how messages carried by
Explaining the regularities of English spelling this code are modified by the contexts in
and the basically morpho-phonemic character which they occur: modern linguistics has
of English orthography helps, but only for taken the microlinguistic approach.
words that follow the rules. Using phonetic Consequently, CA has also taken this
transcription helps, but we can’t expect a approach. There has recently however been
student to run to the pronouncing dictionary increasing attention to contextual
before uttering every new vocabulary item he determination of messages and their
has learned in written form. In the final interpretation, a growing concern for
analysis, we have to correct each macrolinguistics. This is not the place to
pronunciation error as it occurs in the explain this shift of emphasis, but we may
classroom. But if the student is not made point out that it coincides with a growing
consciously aware of the cause of his error in interest in semantics, sociolinguistics,
terms of interference, the next time around he discourse analysis, speech-act theory and
is likely to revert to illogical guesses. ethnomethodology.

The cure for Spelling-to-Sound interference First and foremost, CA owes to linguistics the
lies in strengthening the association between framework within which the two linguistic
visual images and auditory images. We need descriptions are organized. By ‘framework’ we
to design better materials for teaching mean three things. First, CA adopts the
pronunciation that deal with problems of linguistic tactic of diving up the unwiedly
spelling contrasts as well as problems of concept “a language” into three smaller and
sound contrasts. A complete description of more manageable areas: the levels of
phonological performance recognizes the phonology, grammar and lexis. Secondly, use
equal partnership of phonology and is made of the descriptive categories of
orthography in the representation of linguistics: unit, structure, class, and system.
language. It also recognizes that partners do Thirdly, a CA utilizes descriptions arrived at
not always agree. under the same ‘model’ of language. We shall
now consider each of these in turn.
Contrastive Analysis
Levels of language.
Carl James
Imagine meeting an octogenarian who is the
sole surviving speaker of a language. As a
linguist it is your moral duty to preserve some
60
account of this language in the form of a set phonology of a language is somehow ‘basic’
of descriptive statements. Here are some of and merits priority in description. The idea of
the descriptive statements which might be feasibility derives from the fact that the
made: sound-system (phonology) of a language is
more finite, more of a ‘closed system’ than
1. This language (L) uses the sounds /ł/, β/, the grammatical or lexical systems and
/θ/, etc. therefore more amenable to exhaustive
2. L has four words for ‘cousin’, depending description. There is much truth in this:
on whether the cousin is male or female or Stockwell and Bowen (1965: 116) are able to
on your mother’s or your father’s side of say with little fear of contradiction: “Spanish
the family. has nineteen consonants including two
3. L shows plurality of nouns in four different semivowels”. By contrast, no linguist could
ways, each involving addition of a claim to know how many syntactic patterns or
consonant to the end of the noun in its how many lexical items there are in any
singular form. particular language: at best he would hazard
4. To ask a question, take the finite verb approximations. The claim that phonology is
(which is in initial position in declarative somehow more ‘basic’ is less easy to justify. It
sentences) and transpose it to sentence- is true that every utterance in a language
final position. must employ the appropriate phonological
segments if it is to be understood: but
No one of these descriptive statements likewise every utterance has to have some
encapsulates a total description of L, of syntactic structure to qualify as an utterance
course: but the more there are, the fuller the of the language in question. The fact that any
description becomes. Notice that each given phoneme has a greater probability of
statements restricts itself to some aspect of L, occurrence in speech than any morpheme or
and does not pretend to cover several aspects any syntagm is not an index of the basicness
of L simultaneously. So 1) says a little about of phonemes, but of their limited number, the
the sound system of L; 2)says something fact that they comprise a small closed set. It
about its lexical stock; 3) describes an aspect is an undeniable fact, however, that the
of word-formation, or morphology of L; while procedural direction for describing the
4) talks of the arrangement of words in L, the phonology first has been observed by
syntax. In other words, linguistic descriptions structural or ‘descriptivist’ linguists, frequently
are approached observing the principle of to the relative or total neglect of the other
‘division labour’, each statement –or grouping descriptive levels.
of statements- being aimed at one of the
levels of language. The four descriptive Mixing levels
statements of our hypothetical last-surviving
native-speaker is each made on a different The second repercussion emanating from the
level: observance of levels of description has been
the injuction that they should not be ‘mixed’.
1. On the level of phonology. In other words, it was a regulation within
2. On the level of lexis. structural linguistics that the description of,
3. On the level of morphology. say, the level of phonology should be carried
4. On the level of syntax. out without reference to the other linguistic
levels. To invoke grammatical factors to
Procedural orientation facilitate the description of the phonology of a
language or vice versa, was viewed as
Two further points should be made concerning illegitimate and this ‘mixing of levels’ was
the observation of linguistic levels for ruled out of court. Nowadays mixing is
description. First, there has been a traditional allowed, and sometimes found to be
‘procedural orientation’ which has dictated necessary to account for some fact of
that, in the course of producing a total language. Hetzron (1972), for example, in a
description of a language, the phonology has paper entitled ‘Phonology in Syntax’ shows
been producing a total description of a that it is necessary to invoke phonological
language, the phonology has been described factors to explain why, of the following
before the morphology, and the morphology Russian sentences, 1) and 2) are
before the syntax. This ‘direction’ of grammatical, while 3) is not.
description seems to have been dictated by
two things: the linguist’s perception of 1. mat’ rodila doč: ‘mother gave-birth-to-fem
feasibility, and a conviction that the daugther’
61
2. doč’ rodila mat’: ‘daugther gave-birth-to-
fem mother’ UNIT. The units of grammar which enter into
3. etu doč’ rodila mat: ‘this-Acc. Daughter the description of English and any ‘related’
gave-birth-to-fem mother’ language are: sentence-clause-phrase-word-
morpheme. Here they are arranged on a scale
Hetzron concludes (p. 253): “Initial object is from ‘largest’ to ‘smallest’, which implies that
possible when the accusative marker is not any unit consists of one ore more instances of
homonymous with the nominative... The the text lower unit, and vice versa, that any
reshuffling of SVO-OVS is blocked when such unit is a direct constituent of the next higher
a homonomy would result”. ‘Homonomy’ is a unit: sentences consist directly of clauses,
phonological feature, determining, in these clauses directly of phrases, and so on. This
examples, syntactic possibilities: to explain order of direct inclusion in turn implies a scale
why 2) is ruled out one must mix levels. which is called the rank scale.

CA likewise observes the principle of linguistic In traditional CA, as in traditional linguistics,

levels, and in the next chapter we shall be one does not analyze, nor, in the case of CA,
discussing in turn phonological, lexical and juxtapose, units larger than sentences. A
grammatical CAs. Now, any CA involves two single sentence in L1 will always correspond
steps: first, there is the stage of description on an one-to-one basis with a single sentence
when each of the two languages is described in L2: the main difference is that some
on the appropriate level; the second stage is languages have to be more explicit than
the stage of juxtaposition for comparison. In others. Note how explicit English is compared
the first stage the observance of levels can be to Russian in the following translationally-
adhered to, but it will frequently be necessary, equated pair of sentences:
at the comparison stage, to cross levels.
Indeed, the degree to which it is necessary to Ix povitaskal: I’ve finished dragging them out
cross levels at this stage is a useful measure in all directions one at a time.
of the degree of interlingual non-
correspondence (contrast) between L1 and L2. CA is therefore concerned with the possibilities
of, and limitations on, maintaining 1:1
Categories of grammar correspondence of units at ranks below that of
sentence. In the following sentence-pair
Consider again the ‘descriptive statements’
pertaining to our moribund language above. The pupil (who has fallen asleep) is Peter.
Besides restricting itself to one of the levels of Der eingeschlafene schuler ist Peter
the language we may note that each makes
reference to various grammatical entities or The English version consists of two clauses,
concepts. Thus, 1) refers to sounds or whereas the German version is a one-clause
‘phones’; 2) refers to a class of nouns, 3) to sentence: at clause rank there is a 2:1
four different ways of marking a noun as interlingual rank shift is called for. A more
plural, four allomorphs of the morpheme complex set of shifts is exemplified in the
‘plural’; and 4) refers to two sentence-types, following Russian/English pair:
which it differentiates on the basis of the
relative order of their word-classes. In other Ona dočitala etu knigu
words, linguistic descriptions are organized S. Cl. Phr. Wd. Morph.
within a framework of categories. Halliday 1 1 2 4 10
(1961: 247) suggests that there are four such 1 1 2 6 8
fundamental categories: unit; structure, class,
and system. Moreover, these four categories She has finished reading this book
are universal: they are necessary and
sufficient as a basis for the description of any The two sentences are unit-identical
language –which adds to their attractiveness (isomorphic) down to the rank of phrase: now
for the contrastive analyst. Only these four they begin to diverge, the Russian sentence
are required, no more and no fewer: “because employing four words, the English six. This
language is like that –because these four, and imbalance is reversed when the morphemes
no others, are needed to account for the data: are counted for each sentence, as follows.
that is, to account for all grammatical patterns
that emerge by generalization from the data” (Russian): on/a/do/čita/l/a/et/u/knig/u = 10
(Halliday, op.cit.). Let us consider these four
categories in turn.
62
(English): above –determined by the place which the
she/has/finish/ed/read/ing/this/book = 8 element is to occupy in the structure. ‘Choice’
here means “the selection of one particular
STRUCTURE. This category is one of the most term at one particular place on the chain in
familiar to language teachers who have preference to another term or other terms
adopted a ‘structural’ approach. “A structure which are also possible at that place” (Muir,
is thus an arrangement of elements ordered in 1972: 10). For example, we must use a
‘places’ (Halliday, op.cit.: 255). The ‘elements’ nominal class phrase to fill the subject slot in
making up the structure of the unit clause in the clause: but we are free to choose between
English are the Subject, Predicator, a singular and plural nominal phrase. When
Complement and Adjunct as in: ‘The cat(s) we come to the slot P, we must use a verb
caught(P) a mouse (C) last night’(a). A phrase, but we are free to choose between
nominal group such as ‘the green shed past and present tense forms, and
outside’ has the structure DEHQ: Determiner simultaneously between perfect or non-
(the), Epityhet (green), Headnoun (shed) and perfect, as well as between progressive and
Qualifier (outside). Each of which is a word. non-progressive forms in English: there are in
Morphemes, being the smallest units on the English three simultaneous two-term systems
level of grammar, have no grammatical from which choices must be made. Systems
structure, of course: they are composed of operate over the domains of units: there are
phonological units. on the level of phonology systems of sentences, of clauses, of groups,
one would say that the words [strit] and of words and of morphemes. Typical systems
[æktƏ] have the structures CCCVC and VCCV at clause rank are mood, transitivity, theme,
respectively, where ‘C’ means consonant and and information. (cf. Muir, op.cit. 119). The
‘V’ vowel. mood system offers a choice between
indicative and imperative; if the speaker
CAs have traditionally focused on the category selects indicative, a second choice is open to
structure, in this sense of the possible linear him, between declarative and interrogative,
arrangement of units into clauses, phrases, and so on. It is likely that all languages
and words. Typical CA structural statements operate the system of mood: but they are
are implicit in the following: liable to differ in the formal characteristics of
the ‘exponents’ as they are called, of any
My father, who plays chess, is very patient. option chosen.. we know, for example, that
Mein Vater, der Schach spielt, ist sehr the German who chooses simultaneously the
gedulging. imperative option from the mood system and
the polite option from the defense system and
In English relative clauses, the finite verb the polite option from the deference system
occupies second position, before the will commit himself to the exponent Kommen
complement and after the subject pronoun: Sie morgen, which has a PSA structure,
Spron.+Comp.+Vfin. whereas a Frenchman, making the same two
selections from the same two systems (mood
CLASS. There are restrictions on which units and deference) will produce a PA structure
can operate at given places in structures. such as Venez demain.
There is one class of the unit phrase which
can fill the Predicator slot in the clause: this Languages may differ, not in demanding
we call the ‘verb phrase’. ‘Thursday next’ different structural exponents of identical
exemplifies a unit phrase which typically system or system-combination choices, but in
occurs as Adjunct: this we may call an offering different ranges of options. For the
instance of the class ‘adverbial phrase’. An system number we normally recognize two
interlingual class contrast at clause rank is terms in English: singular vs. Plural, whereas
exemplified in: in some languages, like Arabic, there is a third
term, dual. Similarly, English operates a two-
V Londone tumano: London is Foggy system of case, the terms being common and
genitive. In Russian, by contrast, there are six
In Russian, a locative prepositional phrase can cases: nominative, accusative, genitive,
occupy Subject position, but not in English: instrumental, prepositional and dative
• In London is foggy. (Bidwell, 1969: 23): a language like Finnish
uses even more.
SYSTEM. And finally, each language allows its
speakers choices from sets of elements which
are not –unlike the class-choices exemplified
63
Microlinguistic Contrastive Analysis Newmark and Reibel (1968: 161) condemn
contrastivists for their assumption that
In this chapter we shall be concentrating on humans “…learn a new language… one bit at a
the practical matter of executing CAs. We time.” These critics seem to overlook to facts
shall focus on the traditional practice of of descriptive expediency, that it is the
microlinguistic or ‘code-oriented’ CAs on the conventions for standing points of interlingual
three levels of phonology, lexis and grammar. difference which give the false impression that
CA endorses an atomistic view of language
Notice further that the title of this chapter and of language learning. It is impossible to
treats CA as a mass noun. Although I myself say how two complex systems such as
am guilty of not consistently adhering to this languages contrast without first reducing
principle, it is nevertheless one worth these systems to manageable subsystems. As
upholding. The principle is that doing CAs of a Halliday, McIntosh and Strevens (1964)
global and exhaustive nature is neither observe: “There can be no question of, say,
feasible nor desirable. Such CAs are infeasible comparing English and Urdu… One may be
simply because Linguistics is not yet in a able to compare, for instance, the nominal
position to describe a language ‘in toto’, so group of Urdu, or English clause structure with
there are no pairs of total descriptions for Urdu clause structure; but one cannot
input to CA. they are undesirable because it is generalize from these two comparisons.”
inconceivable that a learner could gain access Furthermore, since most CAs are destined for
to, or be exposed to, the whole of the L2 in an eventual pedagogic stage, to prepare the
instant: to suggest that he can is to subscribe ground for the sequencing and grading
to what has been called the ‘blindingflash implied by pedagogy. Even if an L2 is not
fallacy’ (Sciarone, 1970). In fact there has learnt atomistically, that is no reason for not
been no global and exhaustive Ca published to making it available for learning “one bit at a
date. The volumes of the Chicago series time”.
(Ferguson, ed.) carry titles suggesting a claim
to be global, but they turn out to be In what follows, therefore, I shall not
superficial sketches of the major areas of the apologize for failing to demonstrate how full
grammar they set out to describe. The CAs are executed: we shall content ourselves
publications emanating from the various with doing some bits of CA, and pointing to
European CA projects do not even attempt to the general principles guiding this practice.
be global, but consist of anthologies of studies
concentrated on selected areas of the General principles
grammatical, phonological and lexical systems
of the pairs of languages concerned. It is, I Before suggesting how CAs are executed on
think, salutary to think in terms of doing some the various levels of language, it will be useful
relevant bit of CA rather than to set out to do to outline the general principles of the
the CA of two selected languages as wholes. procedures: since repetition will be avoided by
so doing, a measure of economy will be
The practice adopted for CA of executing gained.
partial differential descriptions of selected
systems and structures of L1 and L2 has, Executing CA involves two steps: description
however, attracted criticism. Contrastivists, and comparison; and the steps are taken in
especially those working within the that order. These two procedures cannot be
Audiolingual movement of foreign-language said of characterize CA uniquely, however.
pedagogy, attempted to execute CAs which Indeed, Corder (1973: 144 ff.) sees the whole
would serve the principles of selection and of Applied Linguistics as involving a first, a
grading advocated by that movement: they second, and a third ‘order of application’, and
singled out areas of L1:L2 contrast which talks of description and comparison being the
would present major learning obstacles in the first and second of these. The same view is
early stages, but would become less difficult implicit in the following much-quoted
as the learner’s knowledge of the L2 statement of Fries (1945: 259) claiming that:
increased. For this they have been criticized “…the most effective materials (for teaching
for perpetuating a naïve view of L2 learning. an L2) are those based upon a scientific
Lee (1968: 192) objects to their practice of a description of the language to be learned,
piecemeal CA, and protests “A language is not carefully compared with a parallel description
a collection of separate and self-sufficient of the native language of the learner.” Note
parts. The parts are mutually dependent and what CA consists of: descriptions of L1 and
mutually determinative.” In similar vein L2, and comparison of the two. Furthermore,
64
the two descriptions need to be ‘parallel’. It is possible that T-GG, a product of American
What does this mean? Linguistics, describes English better than it
describes other languages. It seems that
The minimum requirement of ‘parallel Applicative Generative grammar, a model
description’ is that the two languages be devised by the Soviet linguist Shaumjan
described through the same model of (1965) is eminently better suited to describe
description: “im Rahmen der gleichen Theorie Russian, a language with a complex
und mit denselben Notations-konventionen” morphology, than it is to describe English.
(Schwarze, 1972: 20). We shall presently Obviously, distortion would result if we did a
meet alternative models for phonological CA of Russian and English using a model
description. Why, we may ask, must the two which favors one of these languages at the
descriptions be framed in the same model? expense of the other: the descriptions, while
There are several reasons: First, different being ‘parallel’, would be unequal.
models can describe certain features of
language more successfully than other We seem to be faced by a dilemma, then: on
models: T-G grammars can effectively the one hand, there are good theoretical
account for native speaker’s intuitions that reasons for using the same model for yielding
certain construction-types are somehow the descriptions of L1 and Ls; on the other
related (Active and Passive sentences, for hand there are equally cogent practical
example) and that certain others are reasons why this is undesirable. There would
ambiguous; Case Grammars, on the other seem to be two ways out of the dilemma.
hand, provide apparatus for explaining the
semantic affinity between a pair of sentences 1. Describe L1 and L2 data independently,
like using the models which yield the fullest
descriptions of either language, and then
This key opens that door translate these two descriptions into a
And form which is model-neutral. There is a
That door opens with this key. precedent for this Translation Theory,
where use is made of an artificial ‘étalon
Now, it follows that if the ‘same’ data from L1 language’ (Melchuk, 1963: 62) which is a
and L2 are described by two different models, neutral intermediary between L1 and L2:
the descriptions are likely to highlight different in fact it is a composite of the two, or
facets of the data. When this happens, the ‘supralingua’, in containing the features of
subsequent comparison will be unnecessarily both of the L1 construction and of the L2
difficult, and, what is more serious still, the construction. Catford (1965: 39)
analyst will be uncertain of the status of the illustrates this convention in comparing an
contrasts he identifies: are they linguistic English and a Russian sentence which are
contrasts, in representing differences between translationally equivalent.
the L1 and L2 data? Or are they reflections of
the use of two different models, i.e., Note that the English construction selects
description-induced rather than data-induced from the etalon features 1, 3, 5 and 6,
contrasts? It was for this reason that Harris while the Russian selects 1, 2, 3, 4, 5 and
(1963: 3) insisted that comparable 7. these sets of features are those which a
descriptions of two languages will only be good grammar of English and Russian
guaranteed if identical ‘methods’ of would specify, but which no grammar of
description are used for description of the either language would generate all seven
two: “since any differences between these of.
descriptions will not be due to differences in
method used by the linguists, but to Sentence (E) Features in the Etalon Sentence R
differences in how the language data
I 1 speaker ja
responded to identical methods of 2 female
arrangement”. have arrived 3 arrival prishla
4 on foot
Linguistic typology tells us that human 5 anterior
languages fall into several types according to 6 current relevance
7 completed
which grammatical, phonological or lexical
features they show preferences for. If some
2. A second solution would be to abandonthe
models are better at describing certain
requirement that the two descriptions
features, it must follow that some models will
need to be equally exhaustive, or, to use
describe certain languages better than others.
Halliday’s term (1961: 272) ‘delicate’. A
65
 number of contrastivists have suggested realizations it follows that each of its
that a CA should indeed show a statements has very broad coverage of
descriptive imbalance, in favor of the L2. potential utterances. We shall now illustrate
Sciarone (1970: 126) points out that “If how CA utilizes parallel description and
both languages are described beforehand, comparison of types in L1 and L2.
too much, i.e., superfluous work is done
for the sake of CA”. He suggests that less Grammatical CA
attention needs to be paid to the L1 than
to the L2, since it is the latter which must Grammatical CAs are carried out on
be learnt. Slama-Cazacu (1971) suggests comparable systems of the two languages
a ‘procedural adjustment’ of CA which she concerned. In the following example I shall
terms ‘contact analysis’: we should be attempt to produce a step-by-step algorithm
more concerned with what the learner (Levelt, 1970) for the execution of a
does with the L2 than with linguistic representative CA.
knowledge (the L1) he enters the learning
situation. Filipovic (1975: 15) openly Copula Sentences Designating Profession in
asserts that his CA of Serbo-Croatian (L1) English and Brazilian Portuguese
and English (L2) has been descriptively
biased toward the latter. This data is taken from Di Pietro (1971). The
steps involved are:
So much, then, for description. Let us move to
the second step, comparison. Here again we STEP: assemble the data exhibiting the
encounter a number of theoretical problems, relevant systems in each language, viz:
mainly surrounding the issue of criteria for
comparison, or the tertium comparationis:we English Brazilian Portuguese
postpone our discussion of these to chapter 7, He’s a teacher Êle é professor
and concentrate here on how to compare He’s a good teacher Êle é un bom professor
rather than on what basis to compare. They are teachers Êles são professores
Admittedly, this is a somewhat arbitrary They are good Êles são uns bons
approach, since the ‘how’ and ‘why’ are teachers professores
inextricable.
There are in fact several grammatical
We compare ‘types’ rather than ‘tokens’: that phenomena (systems) manifested in each of
is: to refer again to Catford’s example above, these sentences, in addition to the system
we do not compare these two sentences as that is the focal point of the present CA: the
strings of sound or graphic substance, but systems of personal pronouns; the existential
their structures. Their structures are: copula, be in English, corresponding to a
choice in Portuguese between ser and estar;
Pronoun + 1st person + sing - Auxiliary – Past, Participle
the system of word order in the clause, which
happens to be Subject + be + Complement in
I have arrived
each language. We disregard these other
Pronoun + 1st person + sing – Prefix + verb + perfective systems for the moment and concentrate on
+ Past + feminine
the article and modifier systems operating
over the Complement noun phrase in such
Ya prishla sentences.

Any structure, being an idealization, Notice that the English and Portuguese
represents an infinitive number of possible sentences are translations. While it is a
realizations: if the structure is a sentence, it is procedural convenience to work with
the basis of many utterances, as Lyons (1968: translationally equivalent sentences, it is not
176) points out. He explains the difference by necessary to do so: obviously the same
reference to de Saussure’s famous distinction grammatical systems would have been
between parole and langue: “Utterances are brought into play if the English sentences had
stretches of parole produced by native been about ‘a skillful engineer’ and the
speakers out of sentences generated by the Portuguese about ‘um bom professor’. As we
system of elements and rules which constitute observed earlier, CAs aim to be generalized
the langue.” statements about systematic correspondence,
and we should bear in mind continually that
From the premise that CA compares abstract the utterances in the corpus are merely
elements rather than their concrete concrete representations of the underlying
66
 regularities. An obvious danger of working plural noun; and zero before any unmodified
with translation equivalents is that of chance noun, irrespective of the noun’s gender or
correspondence (or non-correspondence) number.
being mistaken for the norm. for example, a
French / English CA based on the translation- The adjective is stated as follows:
pair a pretty girl / une belle fille would lead to
the erroneous generalization that attributive English Portuguese
adjectives occupy pronominal position in both 2) adjective base form/-N adjective
languages, which is manifestly untrue.
2E states that the form of the adjective is
STEP 2: for each language, state the invariant in English, irrespective of the
realizations of each grammatical category number of gender of the head noun. 2P states
pertinent to the CA being done. In the present that the adjective has two realizations in
instance, the pertinent categories are: Portuguese: bom before a masculine singular
indefinite article and attribute. This means noun, bons before a masculine plural.
that in each of our two languages these two
categories accompany the predicate head Notice the caution with which these
noun in sentences identifying individuals by descriptive statements are made: they are
profession: this is the constant across the two accounts of the data upon which they are
languages. Since we are concerned with based, and do not transcend it. This is why in
differences rather than constants, we are as both Portuguese rules, we take the trouble to
contrastivists o0n the lookout for any co- specify that the nouns involved are masculine
occurrence restrictions imposed by either in gender. At this point we say nothing about
language on the ways in which the two the forms articles and adjectives assume in
categories are realized. As we shall see the context of feminine nouns, simply because
presently, the variant realizations of the there are no feminine nouns in the corpus.
category ‘indefinite article’ are determined by This points to the third step for the
two factors: whether an attributive adjective contrastivist to take:
co-occurs in the NP, and whether the head
noun is singular or plural. STEP 3: Supplement the data: since our
interest has been aroused for the ways in
Although this is essentially the descriptive which feminine head nouns in such sentences
phase of the CA, it will be convenient to in Portuguese influence the forms of the
anticipate the third, contrastive, phase by article and adjective, we add two further
listing the descriptions in two parallel sentences to our corpus, together with their
columns. Each statement made at this stage translation equivalents:
is a ‘rule’ in the sense of being the explicit
formulation of a regularity of the language. English Portuguese
She’s a kind nurse Ela é uma enfermeira bondosa.
Rules
They are kind nurses Elas saÕ (umas)
English Portuguese enfermeiras bondosas.

Indef. Indef. Having done this, we perceive the need to

a/ -(adj) N. Sing Um/-adj N. sing. M
¥/-(adj) N. Pl. Uns/-adj N. pl. m
expand the Portuguese rules to accommodate
Article Article
¥/-N the new data. The reader is invited to rewrite
rules 1P and 2P. we move on to the fourth
and final stage of the CA, namely:
Rule IE (English) states that the indefinite
article is realized in one of two ways: as a STEP 4: Formulate the contrasts which have
before a singular pronoun, irrespective of been identified by the analyses of Steps 2 and
whether that noun is premodified; and as zero 3. this then, is the contrastive phase proper;
(ø) before a plural noun, premodified or not. and it is where we face a number of
The brackets on (Adj) are a convention for procedural problems. The foremost of these
stating that the adjective is optionally present. concerns the formulation of contrast: whether
contrast is best stated in terms of imbalanced
Rule PI (Portuguese) states that the indefinite equations or in terms of operations. It is a
article has three possible realizations: um decision which will, in the main, be dictated by
before a singular masculine noun modified by the ‘model’ of grammatical description one
an adjective; uns before a modified masculine has chosen. The decision that has to be made

67
we can conveniently be referred to in terms of than refer to the Portuguese variants as un,
Hockett’s (1954) distinction between IP (item- uns, etc., it might be preferable to use
and-process) and IA (item-and-arrangement) subscripts and refer to them as ind. Art. Port.,
models for monolingual description. We have ind. Art. Port., etc.
also mentioned that Harris (1954) in an article
entitled “Transfer Grammar” nominated the IP The A approach Eschews the task of producing
model for comparative purposes. Harris’ algorithms for converting a grammatical
suggestion is that it is possible to formulate a system of one language onto that of another.
set of instructions which, when applied to the Instead it states the relationship in the form
grammar of one language, will yield the of a set of equations. Although this approach
grammar of another. Let us consider what lacks the dynamism implied by transfer rules,
form these ‘instructions’ would need to take to it is preferable for other reasons to be
deal with our Portuguese / English data. discussed below, and is in fact the approach
anticipated by our ‘parallel’ descriptions of the
For the indefinite article we start from the English and Portuguese data we have been
position that English allows the option examining. The equational representation of
between overt a and zero, the choice being pertinent contrasts might take the following
determined wholly by whether the head noun form:
denoting profession is singular (a) or plural
(Ø). To show the relationship between English English Portuguese
Ø and Portuguese Ø we have to add two A/- (Adj) +N. sing. Um/-Adj + N. sing. Masc.
Uma/-Adj + N. sing. Fem
instructions to the ‘transfer grammar’. The
Ø/-N. sing
first is to relax the singular vs. plural Ø article/ -(Adj) +N.pl. uns/-Adj + N. pl. masc
condition; the second is to introduce a umas/-Adj + N. Pl. fem.
condition that the head noun may not be Ø/-N. Pl.
premodified. The two transfer (TR) rules will
therefore be as follows: There are three things to notice about such
equations. The first is that, being ‘statistic’
TR1. Indef. Article Ø/-N (sing.) (Pl.) M accounts, they can be read in either direction:
left-to-right and vice versa. Transfer rules, by
We introduce a convention of including within contrast, are inherently directional: the rules
a box labeled ‘M’ (for modification) the crucial describing the conversion of English into
feature of the transfer rule, i.e., the feature Portuguese are different from those effecting
that carries the specific contrast. the conversion in the opposite direction.
Secondly, the equations deal with concrete
TR2. Indef. Article Ø/- Adj M N. phonological realizations of the category of
indefinite article in the two languages. While it
Here again the crux of the contrast appears in makes little sense to talk of converting English
a box labeled ‘M’ to indicate that for Ø to /∂/ into Portuguese /υm/, as we have seen,
occur the Portuguese noun must not be there is no objection to equating these
premodified: note the minus sign. phonological strings in the two languages. And
thirdly, the equational statement allows one to
Similar transfer rules will have to be see at a glance which language has the
formulated to introduce Portuguese-specific ‘richer’ or more finely differenciated set of
conditions for the overt realization of the realizations (system) of the relevant category.
indefinite article as um, uns, etc. note In our example we see that there are no fewer
however that it is not the task of the transfer than five terms (um, uma, uns, umas, O) in
rules to specify the real phonological values of the Portuguese system, corresponding to the
these alternative realizations of the article in unique term (∂) in English. This fact of
Portuguese: in other words, there is no interlingual multivalence has implications for
question of rules converting English /∂/ or /εI/ learning.
to /υm/ or /υnz/. As Makkai says (1971: 168):
“…the transfer rules do not need to tell me the At this point, with the explicit statement of
specific phonological shape of the form interlingual contrast, the CA proper is
transferred to. This is derived from the complete.
structural description of the language itself.”
It is not a matter of converting /∂/ to /υm/, Phonological CA
but of specifying how a grammatical category Contrastive Phonetics and Phonology
of English gets parceled out as a
corresponding category in Portuguese. Rather
68
In the previous section I said that function? Indeed they can, by taking as the
grammatical analysis concerns itself with criterion for comparison the articulatory grid
types rather than with their physical employed in the IPA chart: on this articulatory
manifestations or tokens. In other words, the framework he can compare similar sounds of
grammarian studies the functional patterning L1 and L2 and match them as being both e.
of classes of linguistic units, not individual g., ‘labio-dental fricatives’ or ‘half-close
words and morphemes as physical entities. A unrounded vowels’. These feasibility of this
similar distinction can be drawn between the approach is guaranteed by the fact that the
role of the phonetician and that of the world’s languages do tend to employ sounds
phonologist. The phonetician is concerned produced by a limited number of combinations
with three types of physical reality when he of articulatory features. This is not surprising
studies the sounds of language: in view of the fact that man’s vocal apparatus
is physiologically uniform throughout the
1. “He is interested in the way in which the world: “Perhaps the most interesting fact
air is set in motion, in the movement of about the pronunciation of language in
the speech organs… This whole area of general is that there are enormous
interest is generally known as possibilities in the number and variety of
articulatory”: phonetics” (O’Connor, 1973: sounds that the human vocal apparatus can
16). produce, and yet only a small fraction of this
2. “He is interested in the way in which the potential variety is actually put to use in
air vibrates between the mouth of the natural languages” (Stockwell and Bowen,
speaker and the ear of the listener… This 1965: 3). The first approach to phonetics CA,
is the domain of acoustic phonetics” (ibid). therefore, is in the comparison of L1 and L2
3. “He is interested in the hearing process… sounds with a shared articulatory basis.
in the sensation of hearing, which is brain
activity… This is the domain of auditory A second approach is physical rather than
phonetics” (ibid). physiological, and is associated with the
acoustic properties of speech sounds. If we
Now speakers of the same language may compare the initial consonants /p/ in the
speak with different accents, these differences French word pale and the English word pal, we
being attributable to different regional, social, can establish that the English plosive in this
or even purely idiosyncratic conditions, and it initial position is accompanied by a puff of
is the phonetician’s task to identify and breath or ‘aspiration’, which is not true for the
classify these variations and to specify their French plosive. While the differences can be
range. At this point the phonologist takes over traced to an articulatory source it is more
–although, of course, the phonetician and easily demonstrated and described in physical,
phonologist may well be one and the same acoustic terms. There are even instruments,
person. The phonologist, however, is such as the sound spectrograph, which record
concerned not so much with the finer details the occurrence of such aspiration. Similarly,
of phonetic variety as with the functional there are acoustic differences, which can be
identity, as tokens of a type, of these demonstrated instrumentally, between the
variants. As a ‘functional phonetician’ he is ‘similar’ vowels in English spleen /splin/ and
interested in “the way in which sounds a German spiel /Spil/ ‘game’. An acoustic
function in a particular language, how many approach to phonetics CA consists therefore in
or how few of all the sounds of language are comparing L1 and L2 sounds that have much
utilized in that language and what part they in common physically and noting the
play in manifesting the meaningful distinctions differences accompanying this similarity.
of the language” (O’Connor, ibid.).
The third type of phonetics is auditory
Such a division of the phonetic sciences into phonetics: it is concerned with what ‘message’
these two main branches immediately poses a the ear transmits to the brain. To take a
problem for the contrastivist: is he to do simple unilingual example: it can be shown
Contrastive Phonetics or Contrastive that the first and second consonantal
Phonology? The former will involve him in segments in English /pit/ and /spit/ are
making detailed descriptions of the sounds of different: in the former /p/ is aspirated, but
a pair of languages and then somehow not in the latter. Nevertheless, the English ear
equating certain of these sound interlingually does not send to the English brain any
for purposes of comparison. But can such instruction to register this phonetic difference:
equations be made pre-phonologically?, i.e., auditorily, and mentally, /p/ and /p’/ are
without reference to the differences in perceived as the same phoneme /p/. notice
69
that we are now speaking of two allophones variants (allophones) for L1 and L2; state the
being tokens of the same type, as having distributional restrictions on the phonemes
equal functions in the economy of English. We and allophones for each language. By the
are no longer concerned with physical or large, the literature on phonological CA shows
physiological reality, but with mental reality. a large measure of agreement on these four
Our domain is now functional phonetics, or steps, as we shall see, although there are
phonology. Although we have illustrated this differences in terminology, and Stockwell and
principal intralingually, it applies equally Bowen (1965: 5-6) like Burgschmidt and Gotz
cogently interlingually and is the foundation (1974: 197) add a fifth step: a statement of
for phonological CA. the frequency of each phonemic contrast
within L1 and L2. Stockwell and Bowen point
Consider the laterals of English and Russian. out that there are many minimal pairs within
Each language has two lateral sounds: the English, exploiting the phonemic contrast
‘clear’/l/ and the ‘dark’ /ł/ of English are both between /p/ and /b/, whereas there are only
alveolar laterals but /l/ is produced with very few centred on the contrast, between /Z/
simultaneous higher raising of the front of the and /dZ/: pleasure / pledger, lesion / legion,
tongue than of the back, while /ł/ has the etc. the latter contrast has a low functional
opposite configuration. /l/ occurs before load. One might object that such intralingual
vowels and elsewhere, i.e., before consonants contrasting is excessively time-consuming,
and finally. Russian has two laterals also: /ł/ since one has to take every possible pairing of
and /ļ/, the former velarized, the latter the phonemes in the inventory, and that the
palatalized. /ł/ “is a lateral fricative, usually comparison they make between /p/: /b/ and
voiced, with mid-tongue depressed, resulting /Z/: /dZ/ is arbitrary, since while the first pair
in a ‘dull’ ‘hollow’ sound of low tonality, contrast by the feature of voicelessness vs.
something like /ł/ in English bull” (Bidewell, voice, the second contrast does not hinge on
1969: 2). There is ample justification, in the same feature: fricative /Z/ is compared to
Bidewell’s account, for equating the Russian an affricate /dZ/. a more systematic contrast
and English laterals on both articulatory and would be the voiced / voiceless pair /Z/: /S/.
acoustic grounds. But what is the functional Indeed, the /Z/: /dZ/ contrast may be in
status of each? For the English speaker /l/ and English a case of free variation, as in
/ł/ are allophones of the same phoneme, in /gQra:Z/-/gQra:dZ/ as alternative realizations
that each sends the same ‘message’ to the of ‘garage’. Burgschmidt and Gotz make a
brain, namely that in either case the /l/ better case for the absolute relative
phoneme is being used. This can be tested by frequencies of L1 and L2 phonemes being
intentionally switching the clear and dark stated in the CA. They quote Delattre’s (1965:
variants with a word: to the English speaker, 95) frequency-count of the occurrence of the
/łIp/ is still lip and /fIl/ is still fill, and when consonantal phonemes in English and
he hears an Irishman say /fI/ mai glQs/ for RP German.
/fIł maI gla:s/ the message is clear. For the
Russian /ł/ and /ļ/ have different status by We shall now consider each of the four
signaling differences in meaning: /dał/ means proposed steps in turn:
‘he gave’ while /daļ / means ‘the distance’.
Bryzgunova (1963: 83) gives lists of ‘minimal STEPS 1 AND 2: INVENTORIZE THE
pairs’; to illustrate the phonemic status of the PHONEMES OF L1 AND L2
/ł/: /ļ/ contrast in Russian. The important
point to be made in this context is that This first, descriptive, step is not really part of
objectively similar sounds of two languages CA. in fact, for most languages a phonemic
can have different functional statuses; in L1 inventory will already have been made
as ‘the same’, while in L2 the same objective available by the phonologist. The
difference is upheld as constituting a contrastivist’s task consists in equating
functional difference. This contingency is the phonological categories across the two
cornerstone of contrastive phonetics and languages. I have already suggested that the
phonology. categories of the IPA chart can be adopted for
this purpose. The consonants of L1 and L2 can
Contrasting sound systems conveniently bre classified according to the
place and manner of articulation and placed in
There are four steps involved in executing CA the appropriate cell of the chart, with the
of the sound systems of two languages: draw voicelees / voiced pairs (e.g., /p/: /b/)
up a phonemic inventory of L1 and L2; equate appearing in this order consistently. IPA
phonemes interlingually; list the phonemic symbols can be used to represent the sounds.
70
 For the vowels, the conventional vowel- system of which is “the essence of simplicity
diagram can be used, which allows a and elegance” consisting of only five pure
specification of any vowel according to the vowels (Stockwell and Bowen, op. cit.: 73).
tongue position during articulation. Rounded
or unrounded variants can be inserted in The question immediately arises as to whether
brackets, and there are diacritics available to the segments entered into these charts are of
indicate any special extra features, such as allophonic or phonemic status in the language.
nasality (-) or length (:) It has been my own In practice we rely heavily on the criterion of
practice to use unusually large charts and minimal pairs: we mentally search the lexicon
diagrams to cater for double entries (of L1 for pairs of words that are differentiated by a
and L2), and I use different colored pens to single phonological segment. This is what the
write in the sounds of L1 and L2. a further Portuguese students did: in establishing the
possibility, suitable for classroom status of /b/ for example, in /bata/, /beNtu/
demonstration of contrasts, is to use two and /bElA/, the following contrasting lexical
transparencies, one being superimposed on items were cited: /patA/, /veNtu/, /gEäa/ etc.
the other for overhead projection. Two vowel likewise for the vowels, minimal pairs like
diagrams may be used, one for monophthongs /sed/: /sEd/ and /ә’vo/: /ә’v/ were cited. The
the other for diphthongs. The following two allophonic status of /l/ and /É/ was
figures illustrate how a class of Portuguese established by noting that /l/ occurred word-
teachers handled the inventories of the initially and medially, but not finally, where
consonants and the pure, nonnasal vowels of /É/ occurred. This brings us to the next step in
Portuguese, using an adaptation of IPA charts: the CA.

Manner Plosive Nasal Fricative Affricate Lateral Vibrant STEP 3: STATE THE ALLOPHONES OF EACH
VCL/ VCL/ VCL/ VCL/ VCL/ VCL/
Place
VCD VCD VCD VCD VCD VCD
PHONEME OF L1 AND L2.
Bilabial Pb m
Labio-
fv We have already seen examples of this
dental
procedure: the aspirated and unaspirated
Denti-
td n sz l pairs /p’, p; t’, t; k’, k/ occurring in English,
alveolar
Palato-
ø SZ tS but not in French. Another example was the
alveolar
allophonic variants of the lateral phoneme in
Palatal cj ´
Velar Kg N t English, as contracted with the phonemes /É/
Uvular R and /ļ/ in Russian. Politzer (1972: 129) has
Apical r identified a number of ways in which pairs of
languages can exhibit contrasts over the
respective statuses of their speech sounds:
CLOSE
a) For two equated phonemes, one of L1 and
O/ o / u/ one of L2, allophonic variants occur for
O/i/ o / U/ one but not for the other. For example, we
equate the laterals /l/ of German and
O ə English. We now discover that the German
O/e O/α BACK lateral is always realized by a ‘clear’ /l/
FRONT O/E while in English there are two allophones
O / o/ in complementary distribution. The
German thus says /liNks/ for links ‘left’
and /fɔl/ for voll ‘full’, while the
O / a/
Englishman says /liNks/ for links and /fUÉ/
O //
for full.
OPEN b) What is an allophone in L1 is a phoneme
in L2, where the sounds concerned are
I do not pretend that these ‘analyses’ of physically very similar. Our example of
Portuguese are either complete or this type of contrast was the ‘clear’ /l/ of
uncontroversial (Strevens, 1954). They are English, equated with the palatal /ļ/ of
merely illustrative of what students with a Russian: the former has allophonic status,
bare minimum of linguistic training can the latter phonemic status.
produce in the framework I am proposing.
Moreover, Portuguese happens to be a In Portuguese the phoneme /d/ has two
phonologically highly complex language, allophones: it is realized as /d/ word-
unlike, for example, Spanish, the vowel initially /dalia/ dália, ‘dahlia’, after a
71
 consonant /alda/ Alda ?girls name’, and phonemes with phonetically very similar
before a consonant /adrianu/ Adriano allophones, but where the environments for
‘Hadrian’. In intervocalic position, these allophones are not identical. Both
however, it is realized as a dental voiced Spanish and English for instance have the two
continuant /D/. This sound is physically sounds /n/ and /N/. The former, /n/, occurs
like the English /D/ in /Den/ ‘then’, and so before vowels and dental or alveolar
we equate them. However, /D/ has consonants as well as word-finally in both
phonemic status in English but allophonic languages. But the environments determining
status in Portuguese. the occurrence of /N/ are different in Spanish
and English, according to Stockwell and
In fact, category b) could be conflated Bowen (op. cit.: 62). In English /N/ occurs as
with category a): instead of saying that an allophone of /n/ before velars, as in
the fricative is phonemic in English but /sinNk/, /lANIst/. In Spanish it occurs before
allophonic in Portuguese (or Spanish) we segments which Stockwell and Bowen
could have said that English /d/ and /D/ designate as /h/ and /w/: /estraN’hero/
have no allophonic variants, while estranjero ‘foreigner’, /na’raNha/, naranja
Portuguese /d/ has. ‘orange’; /saNwiS/ ‘sandwich’ and /uN’weBo/
un huevo ‘an egg’. This phenomenon, the
c) This category pf contrast applies to pairs contrastive distribution of phonetically similar
of L1 and L2 sounds that stand in a one- allophones, is probably the most formidable
to-one relationship, not the one-to-many one that faces both the contrastivist and the
relationship characteristic of category b). foreign-language learner.
here, the two equated segments have
different absolute statuses in their The relative absolute distribution of equated
respective phonological systems. phonemes of L1 and L2 is a less complex
analytical problem. Although Briére (1968)
There are good reasons for assigning suggested the syllable to be the proper unit
phonemic status to the German palated within which to conduct distributional
fricative /C/ in words like /iC/ ich ‘I’ and investigation for CA, most contrastivists have
/SprEC«n/ sprechen ‘to speak’. This sound continued to take the word as the relevant
occurs in English for some speakers, but unit: so we speak of sounds occurring in
only in word-initial position: /Cju:dZ/ word-initial, medial, or final positions. A
‘huge’ and /Cju:/ ‘Hugh’, where it is phonemic distributional restriction familiar to
obviously an allophone of /h/. it is not, most British teachers of French concerns /Z/
however, a positionally conditioned in the two languages. In French it can occupy
allophone of /h/, since /hju:/ and /hju:dZ/ all three positions within the word: compare
are possible, indeed predominant /Zon/ ‘yellow’ /leZe/ ‘light’ and /g∂RZ/ ‘throat’.
pronunciations. /C/ and /h/ are in free In English /Z/ occurs only medially and finally
variation in this position. They are optional as in /meZ∂/ ‘measure’ and /ru:Z/ ‘rouge’.
variants, the selection of one or the other Consequently, the English learner of French
not being determined by the phonological will have difficulty with the pronunciation of
rules of English. I consider this type of French words having /Z/ initially. For similar
interlingual contrast in the status of reasons, he will experience difficulty with
speech sound to be important: unless a German words having /C/ in medial and final
rigorous phonetic CA is carried out, there positions, even though he has initial /C/ as an
is the danger of overlooking the fact that a allophone of English /h/ in a few native words.
learner of an L2 may have available in his
L1 serviceable sounds of such peripheral We have just discussed what I called
status. ‘absolute’ distribution of sounds. Another type
of distribution contrast concerns the
STEP 4: STATE THE DISTRIBUTIONAL combination of sounds: one language may
RESTRICTIONS ON THE ALLOPHONES AND permit certain sequences of sounds at one or
PHONEMES OF L1 AND L2. another position in the word. This is what I
called the phonotactics of the language.
We already embarked on this operation, when Contrastive phonotactics is an important part
we identified the allophonic variants in the two of phonological CA. in Polish the combination
languages. What is called for now is a detailed /StS/ occurs in al three positions, as in
and fully explicit account if the environment in Szczeczin ‘Stettin’ (a town), jeszcze ‘still’ and
which typical allophones occur. It is possible barszcz ‘beetroot soup’. This phonotactic
for the two languages to have corresponding sequence is impossible in English, although it
72
is possible to find it distributed across a word ‘decisions’ on the part of the speaker (and
boundary as in /freS’tSI:z/ ‘fresh cheese’. Oft- learner), they distinguish optional choices
quoted is the absence, in Spanish, of English from obligatory choices: optional or free
consonant sequences pr clusters like /s+[p-t- choice exists where the speaker selects a
k] + r/ in words like spray, stray and scream; phoneme, i.e., decides whether to say /pin/ or
while the word-final clusters /[n-l] + d/ as in /bin/; obligatory choice or non-choice is when,
world or sound of English are alien to Spanish. having selected the phoneme /p/, he is
constrained by the environment it occurs in to
So far we have restricted our observations to select one of its allophones. Thus the
the segmental phonologies of the two taxonomic models does provide an interesting
languages under CA. of equal or greater and plausible hypothesis about relative
importance is CA of the suprasegmental difficulties of pronunciation. Admittedly, as
phenomena: the features of stress / rhythm Kohler points out, the predicted relativities of
and intonation in particular. Space will not difficulty are not always upheld in practice:
permit us to enter into any details of although English differs from German in
suprasegmental CA. suffice it to mention the lacking initial consonant clusters such as /Sm-
interesting work of Schubiger (1965), who . St-, Sl-/ the Englishman seems to have little
establishes the functional parallelism between or no difficulty in pronouncing such clusters.
English intonation and the German modal
particles, and Zimmermmann’s (1972) A second failing of taxonomic phonology in CA
account of the relationships between is its inability to differentiate productive from
topicalization, word order and intonation in receptive difficulty: it is assumed that what is
the same two languages. difficult to perceive by the learner will ipso
facto be difficult for him to produce. Such is
Phonological models not the case. Examples are legion of an
asymmetry between the learner’s receptive
The range of models available for syntactic and productive control of phonological
analysis is large. For phonological analysis we segments. The English speaker may hear the
have a two-way choice between taxonomic /k/: /x/ contrast between German /lkə/ ‘loose’
phonology (the model which we have been and /lx/ ‘hole’ but be unable to produce the
using throughout this section) and generative /x/. Kohler claims that “/º/ is extremely
phonology. The question inevitably arises of troublesome to produce for most speakers,
the relative merits of these two models. The but very easy to detect” (op. cit.: 85).
taxonomic approach, as we have seen, has
the aim of ‘setting out phoneme systems, Generative phonology stems from America
combinatorial possibilities of phonemes (Chomsky and Halle, 1968) but is rooted in
(phonotactics) and non-distinctive variations European phonological theory of the 1940s.
of these units in different languages’ (Kohler, like generative syntax, generative phonology
1971: 84). Kohler goes on to say: “it can be assumes that surface-structure phonology is
said that on the whole this theoretical derived from the deep-structure phonology by
assumption works pretty well” (ibid). The means of transformations: “The phonological
main value of the phoneme-and-allophone rules… mediate between the systematic
approach is that it identifies two categories of phonemic level (at which all distinctive feature
pronunciation problem which L2 learners face: information is specified) and the systematic
errors resulting from phonemic asymmetries phonetic level (at which all phonetic
between the two languages, and those information is specified)” (Southword and
resulting from allophonic differences. The Daswani), 1974: 77).
assumption, normally upheld by observation
of learners’ speech, is that the first category This is the first weakness of the model, for
will be the source of more fundamental purposes of CA: the phonological deep
distortions, often leading to unintelligibility structure is assumed to contain forms which
while the second category merely leads to are deleted from the surface representation –
‘foreign accent’ without much impairment of ‘king’ for example is given the deep structure
communication. In fact, it is on the basis of /k®Ng/ with the subsequent deletion of /g/”.
the difference between phonemic and the phonological deep structure not only lacks
allophonic contrasts between English and psychological reality, but seems to contradict
Spanish that Stockwell and Bowen (op. cit.: it. With its postulation of these “quasi-mystical
16) draw up an eight-point scale of underlying forms”. Given the choice between
pronunciation difficulty. Seeing language taxonomic and generative phonology, while
performance as a series of segmental accepting that the latter is probably more
73
powerful for ‘pure’ linguistic purposes, we reliant infants, as well as adults learning an L2
should, as Burgschmidt and Gotz (op. cit.: in the natural setting, are, upon leis for
199) do, opt for the former and weaker, for communication. It is these insights, together
the simple reason that it is more practical and with a renewal of interest among linguists in
concrete. problems of semantics (including lexical
semantics) that promise a heightening of
There is however one element of generative activity in Contrastive Lexicology, which has
phonology, the element it inherited from been relatively neglected as one of the three
Prague School phonology, which has proved branches of microlinguistic CA (Roos, 1976;
useful in phonological CA: the concept of Dagut, 1977). While explicit Contrastive
distinctive features. Distinctive feature Lexicology has suffered from this neglect, one
phonology operates on the assumption that should bear in mind that many of the
the phoneme is not the most convenient unit problems to which it will ultimately have to
for phonological analysis, since it can be address itself have been the concern of
analyzed into a set of phonological scholars in related disciplines throughout this
‘components’ or features, which are more century. In the 1920s and 30s Edward Sapir,
fundamental than the phoneme itself. Thus and B. L. Whorf, concerned themselves with
the English phoneme /t/ is a composite of the the problem of linguistic determinism, a
features /+ voiceless/, /+ apical/, /+ stop/, hypothesis claiming that, since language
which distinguish it from /d/, from the labials determines our perception of reality, and since
/p, b/, from the palatals /tS, Z, S/ and from languages are structured differently, different
the velars /k, g/ and so on. There are two language communities have different views of
obvious advantages in this approach. The first what is, objectively, the ‘same’ reality:
is the gain in economy: whereas a language “Languages have a tendency to ‘impose
may use from 30 to 40 phonemes, it is structure on the real world’ by treating some
possible exhaustively to characterize such a distinctions as crucial, and ignoring others”
language using no more than a dozen (Leech, 1974: 30). The Sapir-Whorf
distinctive features. Further economy is hypothesis, then, views language as the
gained by the binary of distinctive feature determinant of perceived reality. This view of
specifications: the presence (+) and the determinism can, and has been, reversed, into
absence (-) of one and the same feature can a claim that culture is reflected in language:
be used as a classificatory index, sparing the “the language of a particular society is an
analyst the multiplication of categories. The integral part of its culture, and … the lexical
second advantage, of particular interest to the distinctions drawn by each language will tend
contrastivist, is the universality of distinctive to reflect the culturally important features of
features: phonemes, in contradistinction to objects, institutions and activities in the
features are certainly not universal, as we society in which the language operates”
have seen. The universal set of features can (Lyons, 1968: 432). Here we have a two-
thus serve as tertium comparationis for stage view of determinism: first culture
phonological CA. in using it we would be given determines language, and then the language
“a much better chance of making fair determines our view of reality.
comparisons between the systems of one
language and those of another” (O’Connor, The Sapir-Whorf hypotheses seems to have
op. cit.: 210). been a particular source of stimulation for
anthropologists. It is they who have
Contrastive lexicology investigated cultural relativity, and so in doing
have shed much light on matters of
The layman’s misconception of second- semantico-lexical relativity. The two bets-
language learning is that it is purely a matter known areas of endeavor on the part of
of the learner learning the lexical equivalents anthropologists are the studies of color
of L2 corresponding to his L1 words. The categories (Berlin and Kay, 1969) and of
structuralist movement in linguistics, and the kinship terms (Lounsbury, 1956;
allied Audio-Lingual Method, with their Goodenough, 1956). It is in this tradition that
emphasis on the priority of grammatical Kalisz (1976) produced his CA of Polish amd
patterns, tended, in contrast to the layman’s English kinship terms.
view, to neglect the role which vocabulary
undoubtedly plays in the process of A second area in which contrastive lexicology
communication. Recent research on language has been kept alive is that of translation. Here
acquisition –of the L1 as well as the L2- has again cultural barriers to effective translation
redressed the balance, in pointing out how have been in the forefront, notably among the
74
Bible translators (Nida, 1964; Wonderly, question which will be one of our concerns in
1968). Wonderly’s book Bible Translations for the rest of this section on lexical CA.
Popular Use has a chapter devoteed to lexical
problems, of which very many are We must now, however, equate lexicology
illuminating. Spanish cimiento is an with lexicography: the latter is one of several
acceptable translation of ‘foundation’ in some practical applications of the former. Likewise,
countries, and more familiar than the lexicon of a language is not the same as a
fundamento: it must, however, be avoided in (monolingual) dictionary of that language.
Peru, since its use there would lead to Both Nowakowsky (1977) and Leech (1974)
confusion with cemento ‘cement’. Similarly, in emphasize the distinction between a
some Spanish-speaking countries dictionary and the lexicon. Leech (ibid.: 202)
‘mature/ripe’ (from Greek teleios) can only be draws a distinction between the practical
applied to grain and fruit, not to people. dictionary or “reference-book on the living-
Wonderly suggests the need for providing room or library shelf” and the theoretical
expansions in translation in certain cases: ‘inbuilt’ dictionary “which every one of us
“the meaning of ‘to serve’ (douleuein) is carries around as part of his mental
delimited contextually as to the quality of equipment as a speaker of a language” and
service… by introducing words for ‘slavery’ constitutes his ‘semantic competence’. This
and ‘master’ into the context”: definition is in line with our general conviction,
expressed elsewhere in this book that a CA is
Romans ‘so that we Que no estuviéramos ya en a differential account of the monolingual’s L1
6:6 may no esclavitud, sirviendo al competence and the L2 competence which, as
more serve’ pecado como a un patrón.
a learner, he aspires to. It is the task of
(Wonderly, 1968: 10)
contrastive lexicology, therefore, to compare
linguistic accounts, stated within the same
Wilss (1977), in his work of translation theory,
lexicological framework, of the lexical
discusses problems of cultural and linguistic
competence necessarily possessed by
relativity attendant on the rendering into an
speakers of the two languages concerned.
L2 of ‘einzelner fur die jeweilige
This is a large-scale and arduous undertaking,
Sprachgemeinscraft charackteristischer
as we shall presently discover, and not to be
Worter’ (individual words that are
confused with such exercises as the writing of
characteristic of a certain speech community).
bilingual dictionaries, not even if they are
He lists such words as: esprit, patrie, charme,
conceived contrastively, as is the case of the
gentleman, fairness, Sehnsucht, Ostoplitik,
Romanian-English CA Project, which has,
Tuchtigkeit (p. 44).
among other things, set itself the task of
producing a contrastive bilingual dictionary of
And of course, where there are L2 learners
the 20 000 most frequent words in Romanian
and translators, there are bilingual
and English.
dictionaries. This, bilingual lexicography, is
the third area in which a practical concern for,
As on the other linguistic levels, the
if not a theoretical commitment to contrastive
contrastive analysis proper presupposes the
lexicology has been maintained. Any
prior analysis of the lexicons of L1 and L2. To
reasonably good bilingual dictionary bears
quote Leech (op. cit.) again: “The lexicon will
witness to this. Consider the entry under
be considered as an unordered list or set of
hawk in Cassell’s New German Dictionary
lexical entries. A lexical entry, in turn, will be
(1957). Three key-words appear:
considered as a combination of three
Hawk1 Die Falke, Habitch (bird of prey)
specifications: a morphological specification, …
a syntactic specification, … and a semantic
Hawk2 Verhokern, feilbieten (offer for sale) specification”. For several reasons, not the
Hawk3 Sich rauspern (clear one’s throat)
least of which is the enormity of the task
required, I shall not adopt this approach to
We have here, then, a 1:3 correspondence in
lexicology for the conduct of lexical CA.
equating the English and the German lexical
instead of producing an ‘unordered list or set’
items; hawk4 in the sense of a ‘plasterer’s
I shall advocate the preselection of various
tool’ is not included. Derivates of hawk1 such
semantic domains (or fields for the purpose of
as hawk-eyed, hawk moth, hawk’s beard,
delimiting the scope of the CA; and I suggest
though at least as rare as hawk4 are included.
further that the lexical entries identified as
We are already in a position to criticize
belonging to the particular fields selected
Cassell’s, on the strength of this one entry.
should be studied and specified according to
This raises the question of what the ideal
their strictly semantic properties: the only
bilingual dictionary should offer its users, a
75
syntactic information pertinent will be in the these four lexical realizations depends on the
form of statements of the co-occurrence values selected for the variables A, x, B in the
restrictions imposed on particular lexical formula. Such verba dicendi as answer, deny,
items. In fact, while not denying that the etc., are not analysable by this formula, but
lexicon constitutes a highly complex and would call for a more complex one containing
ultimate monolithic system (how else could such further variables as antecedent speech
one use it?), for our present purposes it will act and speaker’s presuppositions. Lehmann
be an advantage to view it as a system of (op. cit.) identifies a number of contrasts
subsystems: these subsystems are the lexical between the four English verbs and their
fields we have mentioned. The view we take German ‘equivalents’ sagen sprechen,
of lexis is, a polysystemic one. erzahlen, redden.

Opting for this approach is not to deny the 1. SAY can have as its grammatical subject a
relevance to our enquiries of the general person, ‘text’ or institution:
principles of lexical design, which we further
assume to be true for all human languages. My mother / The brochure / Scotland Yard
Although each field will have its says…
idiosyncrasies, in terms of the number and
nature of its constituent lexemes, as well as of SAGEN prefers a human subject and
the ways in which they interrelate, these rejects ‘text’:
relationships will be of recurrent types; we are
thus in a position to view language diversity, Ihre Brochure sagt…
and contrast, in the lexicon, against a
background of universal formal constraints. 2. SPEAK refers to the faculty and quality of
Another task we set ourselves, therefore, in oral communication:
this selection is to characterize the kinds of
relationships into which lexical items enter He speaks six languages: he’s a French
with each other, within the same field. speaker

Word fields He speaks well: he’s a good speaker.

The concept of word field, which has received TALK, however, refers to quantity:
much attention in diachronic work from the
German linguists Trier and Weisgerber, was He’s a great talker
introduced for the purpose of delimiting the
lexicon into cohesive subsystems. It has REDEN carries both the qualities of SPEAK
affinities with the thesaurus (cf. Roget’s), and and TALK:
contrasts with the conventional dictionary in
identifying within the lexicon a number of Er ist ein gutter Redner
semantic, cognitive, attitudinal or notional
areas of concern; the dictionary by contrast, is Er redet zu viel
organized on the simple alphabetical principle.
Hartmann (1970) lists word fields that have 3. TELL conveys the fact that the addressee
been studied; these include: OFFENCE, JOY, was given information, was commanded,
VISUAL, PERCEPTION, SOUNDS, FACIAL or was entertained:
EXPRESSION, COLORS, EATING, VERBA
DICENDI, PARTS OF THE BODY, VEHICLES, the smoke told us a new Pope had been
COOKING, ARTIFACTS FOR SITTING, PIPE found.
JOINTS; to name but a few.
He told the kids to make less noise.
An interesting recent CA of verba dicendi in
English and German is Lehmann (1977). The He told her a dirty joke.
verba dicendi constitute a notional class of
verbs, and moreover an intuitively plausible SAGEN corresponds with TELL in its
class. Their function is to refer to speech acts, informative and imperative functions:
the basic semantic conditions for which are
uniformly: A says x to B. more precise Sein Gesicht sagte uns, daB er argelich
specification determines the selection of one war.
member from the class: say, speak, talk, tell.
In other words, the selection of any one of Er sagte der kindern, ruhig zu bleiben.
76
 Whereas the ‘entertain’ function is carried linguistic semantics is the COMPONENT, to
by ERZAHLEN: which we now turn.

Erzahl ‘uns mal eine Geschichte’ Semantic Components

Another recent word-field CA of interest is In the previous section we showed how

that by Bančila (1974) on terms for physical phonemes may be analyzed into phonological
pain in English and Romanian: pain, ache, features. Similarly, lexemes can be shown to
headache, stitch, sting, cramp, heartburn, be composed of semantic features or
twinge, sore, smart, earache, sore throat ‘components’. We stated above that a
would be the list of English nouns. The language using 25-40 phonemes can be
addition of adjectives and participles would economically analyzed at the phonological
extend the list, of course, but limitation by level by reference to about a dozen
grammatical class in this way is one legitimate phonological features. Now a typical native
way of narrowing the field. Since German is speaker has a vocabulary of some 20
better known than Romanian, I shall use 000words. Compare this figure whit that for
German to exemplify the interlingual phonological units and it would seems that the
correspondences. number of semantic components needed to
specify a speaker’s vocabulary will be in the
1. Pain, ache, smart, headache, and sore region of 10 000. such is, however, not the
throat, are all realized by German Shmerz case; it has been calculated that: “The
or Schmerzen, with appropriate surprisingly low number of 17 features (log
modifications. So headache and sore 100 000) would suffice to characterize the
throat are compounded with kopf- and lexical units of a language (or dialect) with a
hals- to give kopfzchemerzen and lexical inventory of 100 000units” (Nemser
halsschemerzen respectively, in both and Vincenz; 1972: 288)
cases with the plural morpheme added.
For smart, Schmerz is adjectivally A clear account of components and how they
modified to give heftiger schmerz, in the are identified is given in Lyons (op. cit.: 470).
singular. He asks us to consider sets of words in
2. Stitch, twinge, sting, prick are all realized English:
by Stitch, with occasional noun
modification, stitch is frequently man woman Child
seitenstich, sting by a noun for the insect Bull Cow Calf
agent: Wenspenstich. ram ewe Lamb, etc
3. Cramp is Krampf(en) while heartburn is
Sodbrennen, a compound consisting of We feel that these triads of words represent a
morphemes indicating boiling and burning. common pattern horizontally, so that we could
set up proportions like:
The first impression we gain of this style of CA
is that its delimitation is somewhat arbitrary. Man: woman: child = Bull: cow: calf
On what objective basis does one select a
word field? We have suggested that it is Both ‘man’ and ‘bull’ are (+male), ‘woman’
identified on the basis of some sphere of and ‘cow’ are (+female), and ‘child’ and ‘calf’
human behavior or human conceptualization. (+immature). Vertically we see further
Even if we accept that this is feasible, and find contrasts: all the first set are (+human), all
that our intuitions about what constitutes a the second (+bovine), all the third (+ovine).
‘sphere of human endeavour’ are inter- The features we have isolated are semantic
subjectively endorsed, i.e., other people’s components. Each lexeme is a complex of
intuitions agree with ours, we have still to such components: ‘lamb’ for example is
solve the problem of what to exclude and specifiable as (+ovine, +young) corresponding
include. We might agree that depression does to the dictionary definition of this item as
not belong to the field of physical pain, but ‘young sheep’ or ‘young gregarious ruminant
what shall we do with lumbago, neuralgia, of the species ovis’.
piles, constipation? Are these not pains but
discomforts? Are they not pains per see but Components, like phonological features, may
causes of pain? These questions become be universals:
philosophical in nature, and Linguistics at least
is in no position to give clear answers to “It has frequently been suggested that the
them. The apparatus that is available within vocabularies of all human languages can be

77
analyzed, either totally or partially, in terms of any comparative-contrastive enterprise.
a finite set of semantic components which are Second, it defines for him that background of
themselves independent of the particular likeness against which the idiosyncrasies of L1
semantic structure of any given language” and L2 stand out, and which sets the process
(Lyons, op. cit.: 472) of interference in motion.

he goes on, however, to criticize the Componential analysis provides the

arbitrariness of component-assignment: why contrastivist with a third vital instrument for
for example, should we differentiate man: his work. This is the semantic feature
woman, bull: cow, cock: hen according to the complex. The English word hand is polysemus,
criterion of sex I.e., (+male) or (+female)? In in having at least four senses:
everyday reality, perhaps we differentiate
man: woman by the clothes they wear or by 1. a part of the arm, with fingers
the length of their hair. To what extent these 2. on a watch or clock
attributes, however, are the reflections of a 3. a person who helps with work
more basic sex difference is another 4. a round of applause
philosophical quandary. Leech (op. cit.: 232)
gives some further depth to this question of To do a CA at this stage would involve merely
the universality of components. He first providing the L2 lexical correspondences, as:
distinguishes formal and substantive
universals: claims for such universals, on the Hand1 = die Hand Hand3 =der Hilfsarbeiter
semantic level, would be: Hand2 = der Zeiger Hand4 = der Beifall

1. “All lexical definitions in all languages are Likewise, a word-for-word CA of German

analyzable as a set of components”. Fleisch with its English equivalents would
(formal) merely register the 1:2 relationship, the fact
2. “All languages have the contrast between that Fleisch is at times translated meat, at
(animate) and (inanimate).” (substantive). other times flesh: “lexeme-to-lexeme
comparison of languages would not be very
Leech argues that belief in 1) is usually taken fruitful” (DiPietro, 1971: 121). What we must
for granted by any theoretical linguist – do is specify the conditions governing ‘at
linguistics is all about formal universals. Most times’.
of the disagreement in linguistics surrounds
the postulation of 2), the substantive Componential analysis allows us to do this by
universals: one need not, as a linguist, claim identifying an intermediate level of semantic
that all languages operate the same contrasts. organization between the components
There are in fact two versions of the themselves and the lexical item: this level is
substantive universals hypothesis. The strong that occupied by the semantic feature
version is that all languages have this or that complex. Each such complex specifies one of
semantic category: and this strong version is the senses of a lexeme, as in the diagram:
manifestly untrue. The weaker version takes
form: “There exists a universal set of COMPONENTS SSENSES LEXEME
unsemantic features, of which every language
possesses a subset” (Leech: 233). Although X S1
this formula could be vacuous, Berlin and Kay Y :
(1969) have shown it to be interestingly true.
They calculated that there are 2 048 possible Z S2
combinations of 11 basic color categories, :
L
whereas, on the basis of their study of 100 K
languages, they found only 22 combinations I
occurring: this suggests very powerful Sn
constraints being imposed by languages on M
the way in which their vocabularies in the field
of color terminology are organized. So there is If L= English hand and S1…S4 are its four
some evidence for the existence of senses, we now specify each sense in terms of
substantive semantic universals of language, its components, these being drawn from a set
evidence which is highly attractive to the x-m. I suggest the following assignments of
contrastivist of course, for two reasons. First, components:
the set of universals provides him with the
tertium comparationis, a vital ingredient for
78
Hand1: (part of body), (end of arm), (for equivalents. The second approach starts with
holding), etc. tentative translations and the subsequent
Hand2: (part of clock), (on dial), (moving)… componential analysis is a check on their ‘fit’.
Hand3: (human), (working), (wage-earning)… Let us illustrate the approach by reference to
Hand4: (human agent), (public appreciation), the field of COOKING in English and German:
(movement)… the field for English has been analyzed by
Lehrer (1969). She regards cook as having
Note that some components are shared by three senses. Its most general sense (cook1)
more than one sense of the lexeme hand: means ‘to prepare a meal’ and this belongs to
hand1, hand3, hand4 are all (human), while the field of household tasks with clean, wash,
hand2 shares with hand4 the component repairs, etc. cook2 is less general and
(movement). contrasts only with bake, i.e., It refers to the
preparation of all foods other than those sold
This approach to lexical CA involves the in bakeries. Cook3 is the most marked sense,
contrasting of all the identifiable senses of and the one which our CA will focus: it
equated lexemes from L1 and L2. we have involves the application of heat in some way
discovered that English hand serves for to food. Lehrer says: “The lexical field covered
German hand as well as for Zeiger, by cook3 can be divided into four main
Hilfsarbeiter and Beifall. Such interlingual categories headed by the lexemes boil, fry,
asymmetry or ‘divergent polysemy’ is a very broil, and bake2 (the specific sense)… these
common source of errors among L2 learners: four lexemes, then, are hyponyms of cook3”
the English student of German is prone to use
die Hand to refer to der Zeiger of a clock. Cook3, is the hyperonym or archi-lexeme of
However, there is an alternative approach the field. Broil is an American English usage
based on a different tertium comparationis meaning ‘to cook directly under a heating unit
than the chance formal identity of German or directly over an open fire’ (Lehrer, ibid.:
hand and English hand: this is the approach 44). It is matched in British English by grill
based on the semantic field. Notice that the and toast.
polysemous hand intersects four semantic
fields, those of PARTS OF THE BODY, PARTS Let us take a subset of lexemes from the cook
OF A CLOCK, EMPLOYMENT, and EXPRESSION field in English and German. First we assign to
OF PUBLIC EVALUATION. Now, it is within, them their semantic components, and then we
rather than across, such semantic fields were shall be in a position to do the CA
semantic homogeneity, and maximum
commonality of semantic components are to C1: C2:
C4:
be found. It is for this reason that Bančila (op. C3: in contact C5:
with with
oven with gentle
cit) is able to specify fully the English and water fat
flame
Romanian terms for physical pain by reference Cook O O O O O
to a mere half-dozen components: (diffuse), Boil + - - + -
(continuous), (sudden), (profound), Simmer + - - + +
(localized), and (physical agent). With similar Fry - + - + O
elegant economy DiPietro (op.cit.: 118) Roast - - + - O
Toast - - - + O
specifies eleven senses of meat and flesh by Bake - - + - O
reference to six components: (human), Kochen1 O O O O O
(concrete), (localized), (animal), (internal), Kochen2 + - - + -
and (edible). Kchen3 + - - + +
Braten - O O O O
Rosten - - - + O
Having specified the L1 lexemes defining a
backen O O + - O
particular semantic field, we proceed to the
CA. we are faced now with a procedural
C1-C5 refer to the five components whereby
option: either we produce an independent
these sets of lexemes can be specified and
specification of L2 lexemes (and senses) for
differentiated. By convention + signifies that
the same field, or we utilize translation
the lexeme is marked by having the relevant
equivalence. The two approaches are in fact
component, - that it is marked by lacking it,
mirror-images. If we adopt the first, a native
and O that it does not apply distinctively one
speaker supplies the L2 inventory, and each
way or the other.
lexeme is analyzed componentially. Then
follows a matching procedure: those L1 and
Note what equations and nonequations
L2 lexemes or senses receiving the same
(contrasts) this analysis reveals:
components are by definition translation
79
Cook = kochen1: both mean to prepare food notion of value is SYNONYMY. In that they
in any of the ways specified by c1-c5 have the same sense, translationally
equivalent lexemes of two languages may be
Boil = kichen2: i.e., in water, on-flame, said to be synonymous. This may appear
rapidly. somewhat trivial, but no longer so if we
consider that “It is of course possible to
Simmer = kochen3: i.e., in water, on-flame, extend the application of the term ‘synonymy’
gently. so that it also covers groups of lexical items”
(Lyons, ibid.:). On these grounds it is possible
Braten is specified positively only by the to equate, intralingually, female fox with
absence of water in the cooking process, all vixen, even mature bovine quadruped with
the other components being non-distinctive cow. One might have reservations about such
(marked by O). now braten can be with fat or equations, since native speakers have to have
without, i.e., dry; one can also braten in the special definitional reasons for using the
oven or on the flame. In fact Bratkartoffeln complex forms; however, in American police
are cooked in a pan, on the flame, with fat, jargon it seems that ‘juvenile Caucasian male’
i.e., they are fried; while ein Rindbraten is freely substitutes for ‘white boy’.
prepared in the oven, without fat: it is roast
beef. In other words, braten is a more general Such reservations do not hold, however, for
term, occupying the semantic space of both interlingual equations or simple and complex
fry and roast. This is the case of interlingual lexical items. It is characteristic of any
divergent generality. To differentiate the interlingual lexical confrontation, whether in
senses of braten we could establish two terms the bilingual, the proficient translator, or the
braten1 (=fry) and braten2 (=roast). Such a L2 Learner, that such simple-to-complex
division could be motivated from within correspondences are set up: they ought,
German –avoiding a ‘dependent analysis’ –if therefore, to be accounted for in a lexical CA.
we introduce more components. The most Kirkwood (1966: 177), in a discussion of
obvious candidates for these components German/English lexical contrasts, mentions
would be selectional features. We would say that while German has a set of simple
that braten1 selects objects like Schniken lexemes for ‘brush’ like Burste, Pinsel, Besen,
‘bacon’, Spiegeleier ‘fried eggs’ while braten2 English has the complex equivalents
selects as objects such nouns as Rind ‘beef’, hair/clothes brush, painting brush, sweeping
Schweine ‘pork’ and Kalbs- ‘veal’. brush. Lipinska (1974: 168) phrases the
generalization thus: “a certain meaning x
Toast selects the same components as rosten, which is expressed in L1 by means of one
giving us Rostbrot for toast (n). however, the lexical item is expressed in L2 by means of
relationship is not always so clear-cut: we more than one lexical items which stand in a
have Rostkartoffeln ‘baked potatoes’, well-defined syntagmatic relation to one
Rostpfanne ‘frying pan’, and rostofen ‘kiln’. another”. Examples are Russian Zavtrakat’: to
have breakfast, polish swad: a smell of
C1-C5 fail to distinguish roast from bake, as burning. It os this possibility of interlinual
Lehrer (op. cit. 45) discovered: her promise paraphrasability which guarantees the
solution is in ‘making roast only a partial feasibility of translation, even in cases where
hyponym of bake’. Once again we can one of the languages has a ‘lexical gap’.
distinguish this pair if we resort to further
specification by selectional features: Another feature of this simple-to-complex
bake/backen lexical relationship connects it with another of
our lexical CA categories: divergent
Select objects composed of flour generality. Where a lexical item of one
(cake/kuchen/geback) while roast/braten language e.g. hand or smell or kochen is more
select animal substances, i.e. meats, as we general than a simple lexeme in the other
have seen above. language, it can be ‘narrowed’ by the addition
of words, so producing a complex lexeme:
We have so far made use of three notions hand (of a clock), smell (of burning),
from semantics: boil, fry, roast, etc., were (langam) kochen. The resultant complexes
called COHYPONYMS of the archilexeme cook; then correspond 1:1 with corresponding
hand and die Hand stand in a relationship of simplexes in the other language: Zeiger,
DIVERGENT POLYSEMY; and braten, in swad,Simmer, respectively. This amounts to a
covering the semantic space of fry and roast, claim that all lexical correspondences are of a
shows DIVERGENT GENERALITY. A fourth 1:1 nature, for example that leicht kochen or
80
langsam kochen correspond to English Second Language Learning
simmer. The objection to this is that while
kochen3 (in the sense of simmer) may be Contrastive Analysis, Error Analysis, and
used in combination with adverbs like Related Aspects
leicht/langsam, native speakers of German do
not usually use the adverbs: they seldom feel Edited by Betty Wallace Robinett and
the need to be explicit about the slow rate of Jacquelyn Schachter.
boiling. It seems to be a general feature of
the use of lexemes in languages that speakers The Linguistic Context on Language
use minimally specific words and phrases: Teaching
only just specific enough to avoid
misunderstanding. The contexts of utterance W. R. Lee
usually convey these ancillary features, so
that in: No language is ever studied in a linguistic
vacuum. The environment of learning includes
1. Schmutzige Wasche muss man kochen. at least the same language or another, and
2. Die Milch kochte, bis sie sahnig wurde. may include several languages, audible and
perhaps visible on every side. When a second
We have a case of kochen2 (boil) in 1) and a language is attempted, it is usually in an
case of kochen3 (simmer) in 2) without any environment of the learner’s first language,
adverbs occurring: one knows that dirty linen acquired as a child, so that all around him are
gets boiled but milk simmered. linguistic patterns whose tyranny he is
struggling to escape. And even in a
The study of such conditions for the co- community using the second language he
occurrence of lexical items, which determine needs, a learner has his home-language
‘the selection of some senses and the habits in combat.
exclusion of others’ (Katz, 1966: 205) belongs
to the study of selectional restrictions. These D. Y. Morgan, in a recent article (English
are regulations for what kinds of lexemes can Language Teaching 10, no. 3 1956), divided
occur in certain grammatical contexts, and English language teaching into three
explain for instance the possibility of John categories of presentation, practice, and
snores and the unacceptability of *The remedial work, and argued that the last of
symphony snores by the fact that the verb these can be well planned only by basing it on
snore requires an animate subject. Selectional an inquiry into pupils’ mistakes.. he is
restrictions are of an all-or-none nature, undoubtedly right on the latter point, and has
which distinguishes them from Firth’s shown too that diagnosis of this kind can be
collocations, which is a statistical notion. The readily made. The merit of such an approach
lexicologist can identify classes of ‘habitual’ is that we get down to the pupils’ level and
collocations such as take pictures, take a see some of the difficulties through their eyes.
walk, sit an exam. The contrastivist will study So doing, we are constantly reminded that it
their counterparts in other languages, is not simple English we teach, but English to
discovering contrastive collocations for the Spaniards, English to Chinese, English to
same senses,, as in Bilder machen, einen Poles, to Nigerians, to Pakistanis, to
Spaziergang machen, passer un examen, and Brazilians, to Finns, as well, of course, as to
the like. pupils of various ages, attitudes, and
capacities. It is clear that English, quite apart
In this short account of contrastive lexicology from the local features which characterize it in
many aspects of lexical contrast have different parts of the English-speaking world,
necessarily been unmentioned: in the context appears variously against various linguistic
of SYNONYMY, for instance, we might have backgrounds. Certain characteristics are
discussed the distinction drawn between thrown into relief in some countries, and other
cognitive and connotative synonymy: in characteristics in others, and this because of
languages there are pairs of lexemes that are contrast with the first language, that which
cognitive, but not connotative, synonyms: the pupils speak at home. For speakers of
freedom/liberty; hide/conceal; angst/furcht. I Serbo-Croat or Czech, English is a language of
hope, nevertheless, to have suggested what several past tenses and puzzling article
seem to me to be the most fruitful areas and usage: but these are not a headache to
procedure for this rather neglected level of Spanish or Hungarian pupils. Among the
CA. problems facing Turkish learners are the
English word order patterns, so different from
81
their own; yet word order is much less of a delayed for a high polish to be put on
stumbling block to the Italians or Dutch. pronunciation. It will necessarily be a little
Speakers of tone languages, such as Chinese, awkward and rough in places, and the risk of
have to pay special attention to the very minor faults becoming a habit has
different English use of voice pitch, while courageously to be taken. This does not at all
some nationalities loose sleep over rhythmic mean that pronunciation cannot methodically
patterns. And although certain features of be dealt with.7 Words and structures which
English are no doubt everywhere fairly easy or have already been introduced furnish the
fairly hard to acquire, the difficulty of each of material for short, regular, and intensive
these also has, in accordance with the pupils’ practices involving isolation and comparison
home language, its varying degrees. In fact, not only of sounds, but of rhythmic and
one country’s linguistic mountain, to be intonation patterns, juncture patterns, and in
patiently climbed, is another’s molehill, to be fact of any devices capable of changing the
lightly skipped over. Mistakes analyses based meaning of what is said. At least five minutes
on adequate material show clearly what is of a forty-minute period can be regularly
most troublesome for the learners concerned spared for speech work of this kind, over and
and thus where they most need support. above incidental correction. A comprehensive
review of the phonetic material is unnecessary
However, it is not only remedial work which and indeed digressive. Attention should be
can be guided thus, but the whole of a focused on the difficult points, and those
language course, and at every stage. Writing which cause little bother may be left, more or
is the obvious basis for analysis, but mistakes less, to look after themselves. And this is
in speaking can be noted too, especially by a where mistake analyses come in. for if these
trained phonetician and with the help of a analyses are based on the speech of enough
tape recorder. A broadly based and learners, and of a sufficient variety of
representative collection of spoken and learners, of the same linguistic background,
written errors, sufficiently classified, may help they enable a teacher to prophesy.
to determine several things –the scope and Forewarned is forearmed. Mistakes can, of
nature of pronunciation teaching, the time course, be dealt with as a course proceeds,
given to practice with certain structures, the without the prior help of any analysis; but it is
time given to practice with certain expressions preferable to know beforehand what is likely
and words, and even the order in which these to happen and so to be prepared for the
structures, expressions, and words are necessity of coping with it. If English /o:/ is
introduced. going to be made too much like English /ou/,
for instance, as when George saw Nora at the
Let us look first at the problem of ball resembles /dZoudZ sou nour« «t D« boul/
pronunciation teaching. It is important during (a common mispronunciation among
early language lessons not to let the Russians), the obvious thing is to be ready to
temperature of interest drop too low. “More make it more like /a:/, and to make in
haste, less speed” is a good enough motto up advance its occurrence in texts used. If
to a point, but we must also remember that a learners are likely to want to put the stress
snail’s pace can lead to boredom. Interest is a and pitch-fall too early in sentences
strong driving force at any stage: its value at demanding a stressed pitch-fall near their
the beginning is enormous. It is necessary to end, such as I shan’t go often thanÍnecessary
push on, to get somewhere and to let the (whereÍoftener than necessary would be an
pupils see they have got somewhere. They unusual pattern), the error can be held at
should be able to answer and ask simple arm’s length is the teacher is aware of this
questions, using the present continuous tense, tendency and prepared to impose the suitable
and deal with simple requests and commands, pattern perhaps by exaggerating it in some
before the initial impetus is exhausted. A new way, on the learners’ melodic ear. If, again, it
current will then carry them along, arising is known that glottal stops may be added to
from interest and pride in their linguistic syllables which should begin smoothly with a
acquirement. To begin with, however, there is vowel sound, more trouble can be taken to
or should be an orderly presentation of
commonly used words in commonly used 7
Nor does it mean that attempts at grading the course
structures. But if oral method is used, the phonetically are valueless. Certainly some of the more
majority of the basic sounds on the language awkward successions of sounds can be postponed,and
will occur within the first two or three lessons. intonations other than the simplest can also wait (as
Progress in the meaningful use of sentences in they do in A. S. Hornby’s Oxforf Progressive English
situations real or contrived cannot safely be London: Univerity press, 1954, for example). However,
it is primarily words and structures which are graded.
82
demonstrate the fact of ‘linking’, and mistakes, basing it on a greater number of
additional phrases to impress the point can be learners. Nevertheless the immediate purpose
prepared. There are many ways in which a was served: I was in a better position to
teacher’s knowledge of the language strains decide how teaching time should be spent.
under which his pupils labor and of the kind of There is no need, of course, to act upon the
mistakes they are liable to make during a results of such investigation slavishly. Many
course or even a single lesson can influence would treat punctuation and spelling mistakes
his attitude and teaching plans. It is chiefly a in the way I did, as on the whole less serious
matter of knowing the pupils better than mistakes with, for instance, articles and
beforehand, at least in one respect,, so that prepositions. But at least one gets a fairly
less time and energy have to be spent on clear picture of what the major difficulties are.
getting to know them at the moment of It may be that with some classes of (say)
instruction. Instead of discovering at the last French or German pupils, literal translations
minute their tendencies to err, and so in a from the home language. Such as We are
sense following the pupils’ tracks, we are here since two hour (=we have been here
ahead of the pupils and thus able to lead them since two o’clock) are not one of the
better. commonest types of mistake: there is no
need, therefore, to prepare for extensive
Teaching Prague University students English demonstrations or repeated practice of the
just after the war, and having then an correct pattern. An analysis of mistakes made
elementary acquaintance with Czech, I noted by some West African learners would probably
down and grouped over twenty-five typical show a need for ample practice of word order
errors in their pronunciation.8 Weekly contact in direct and indirect speech, while Turkish
with about three hundred students enabled learners’ mistakes reveal that word order in
me to make this collection of odd moments as general required particular attention. With
we went along. The grouping was somewhat Slavonic pupils, weaknesses on the articles
rough-and-ready too, yet it could form the and some relative clauses, as well as on
basis of a program of ear and speech-organ comma usage, would all be shown up in their
training, one essential mans to the degree of importance by mistakes analyses,
improvement of language-hearing and an so the teacher is better prepared for
language-speaking skill. teaching the point involved to other pupils
speaking the same first language.
As with pronunciation, so with other aspects
of language. Once an adequate analysis of The behavior of words, and of invariable or
errors has been made, this can and indeed slightly variable expressions, in so far as it
must influence the whole course. About eleven can be separated from the behavior involved
years ago I compiled a record of two thousand in the chief syntactic patterns, is still another
mistakes occurring in Czechoslovak students’ aspect of language. What has been said about
essays. The essay themes were various and of the analysis of other types of error applies
their own choosing, the students (about here too. The frequency with which a word
seventy on this occasion) from whose work occurs in English is not the only thing of
the mistakes were gathered were picked out importance about it: there may be no word in
haphazard, and all the mistakes they made the learners’ language more than roughly
were included. Wishing to apply the result of corresponding, or this word may associate
this inquiry at once, I hurriedly grouped them with others a manner quite different from the
into a few categories, e.g., wrong punctuation English word. Choice of vocabulary with which
(14.4 percent of all errors), misuse or to ‘operate’ the basic patterns, however, is
omission of articles (3.6 percent), misspellings commonly determined more by frequency or
(13.5 percent), non-English constructions and by thoughts of its immediate use in classroom
wrong word order (11 percent), wrong use of situations than by considerations such as
tenses (3.4 percent), and so on,. In the light these, at least in a printed course meant for
of further language study, I should probably use anywhere. Study of the mistakes made by
make a somewhat different analysis of this speakers of any other language in using
material today and enlarge the collection of common English words will suggest the
inclinations they have to fight against. To take
a simple instance, Czechs are liable to say The
This sounds a large number but, even so, little
8
news are good and Christmas are coming –the
enough account was taken of rhythm and Czech words for ‘news’ and ‘Christmas’ go
intonation, perhaps more important for with verbal form commonly felt to be plural.
intelligibility than ‘sounds’. On the other hand, Knowing this, a teacher of English to Czechs
several minor mistakes were included.
83
will be ready to give these words special analysis be based? Do we require to know
treatment, and may perhaps postpone using only the errors which the class we are
them. Word-mistake study helps a teacher to teaching is likely to make? Or is it helpful to
see which English words are best suited for have an analysis of more advanced learners’
the ‘operation’ of patterns: possibly those errors as well?
which give difficulty may be avoided for a tie.
It is plain that a whole series of analyses,
Thus mistakes analysis may influence the based on various stages of achievement from
order in which vocabulary is introduced. Can elementary to advanced, is desirable. Light
they also help to determine the similar shed on the immediate task in hand is of
ordering of structures? This may seem prime interest, and the teacher will look firstly
unlikely. It is mainly a question of the for an analysis of errors made by pupils like
usefulness of those structures, dependent in his own. Analyses made at more advanced
part of their frequency of occurrence in stages, however, can also help, showing as
English and in part of their capacity for they do what types of errors tend to persist
accretion. It is necessary to keep dissimilar and which therefore demand skillful avoiding
structures apart: that is to say, not to action or remedial treatment.
introduce them to the pupils simultaneously
nor until those already taught have has a Such analyses are of use to both
chance to establish themselves in the inexperienced and experienced teachers.
learners’ newly acquired speech habits. The Previous experience with similar pupils may
word order patterns of this is a chair, is this a have given a teacher some of the knowledge a
chair?, and what is this? Are an example. If particular analysis yields; but few have had
we teach this question types too soon, there is experience so broad as to be able to forecast
a risk of muddle, resulting perhaps in what the errors of any type of class, even in a
this is?, a pattern which may be hard to get single language area. Mistakes analyses,
rid of. Nevertheless the risk is slight where in especially valuable to the inexperienced, can
the learners’ language there are similar very also be a great help to those faced with a
commonly used patterns, with the equivalents grade of pupil they have not taught before, or
of this is and is this interchangeable in the pupils speaking different first language. Even
same way. This change in English is then no a teacher of great experience, or one to whom
problem. In German for instance, there is Das a particular type of class is not new, may
ist ein Stuhl, ist das ein Stuhl?, and Was ist welcome a systematic statement of
das?, and in Spanish (Esto) es una silla, ¿es difficulties, well illustrated by examples. It is a
esto una silla?, and ¿Qué es esto?. On the valuable guide when planning a course of
other hand, in Czech there is To je stul and je lessons. Without it a teacher is less aware of
to stul?, but commonly Co to je?. Czech the task which confronts him, and so plans
learners of English thus tend to say What it less effectively.
is?, but this mistake is much less usual among
Spanish and German learners. With the latter Through an examination of learners’ mistakes
we need not, therefore, so carefully space out a teacher may enter more fully into the
these patterns to avoid confusion. Again, environment of teaching and put on, as it
learners from some countries will readily take were, his pupil’s linguistic spectacles. This
to the tense change in converting direct to should enable him to see his way more
indirect speech (He said. “I am coming” –he clearly. An obvious question to ask, however,
said he was coming), while elsewhere this will is whether there is not a more effective
call for cautiously graded treatment. manner of looking at things from the learners’
Nevertheless, on the whole it is not so much viewpoint. As we have already noticed, many
the sequence in which structures are taught learners’ difficulties reflect features of the
that mistake analyses are likely to influence home language. Would a knowledge of this
as the speed with which some of them are language not be a better pair of spectacles to
taught. put on? Are we perhaps not doing things by
halves in examining mistakes only? The
Mistakes in the use of sound, words, and question can at once be restated, since in
structures may thus be usefully collected and most parts of the world the majority of those
examined, and mistakes analyses at each of teaching English will already speak the home
these levels can be applied to language language of their pupils. The problem for
teaching.. we are also concerned, however, them, on the contrary, is whether to bother
with stages of achievement in learning the about mistakes analyses. However, a
language. On what stage should a mistake substantial minority of teachers do not speak
84
the learners’ language, and have to ask so who knew a little French, although an
themselves if they can manage well enough extensive context alone might suffice to give
without it. the meaning: “They arrived three days ago”.
Or take a Czech learner’s “I don’t know
At first sight, it would appear that they can. already”, almost incomprehensible without a
What concerns us chiefly, after all, is the knowledge of Czech, which makes it clear that
encounter between the two languages. the speaker wanted to say “I no longer know”.
Attention should be focused, it seems, on the Lastly, a teacher ignorant of the home
struggle of the second language usages language cannot use features of it as starting
against those of the first. If knowledge to the points in instruction. It is an advantage to be
learners’ language enables one only to proficient at making the sounds, for some of
forecast mistakes, why bother to acquire it? them may be modifiable into English sounds.
Those who speak it, on the other hand, would At other level too it is desirable to know what
seem to gain something from a study of there is or is not to work from. Teaching
mistakes, this involves direct concentration on English tense usage to learners whose
what is relevant and the reduction of language has only one past tense form is apt
guesswork to a minimum, to perplex a teacher unaware of the fact, but
errors in several tenses are seen to be linked
To guess at probable types of error from a by one who is so aware. Translation may be
knowledge of the first language only is, excluded from the teaching method:
without doubt, to take a somewhat far-off nevertheless it is wise to look closely at the
view of teaching problems. Thus if a first use of single past tense forms in varying
language has no final /N/, as in laughing, it is contexts, if only to discover that there is little
a good guess that another nasal may be support in the learners’ language for what one
substituted, as in /’la:fin/. But this is not at all is trying to teach. Again, at word level there
the same thing as seeing that it is substituted, can be a chance similarity between the
and in what positions. If a language has no English and a first language word different
vowel sound close to that in bet or that in bat, meaning, as with cigar and Turkish sigara
but only a sound lying somewhere between (cigarette), clinic and Czech klinika (teaching
the two, it is likely that /e/ will often be hospital), station and Spanish estación
pronounced too open and /a/ too close. Yet it (season); if the teacher knows such things he
is surely more helpful to see what happens in can guard the better against
practice, for other factor may be influential misunderstanding.
too, such as frequency of occurrence and the
nature of the other first language vowels. All Knowledge of the learners’ language is of
such factors could perhaps, in forecasting practical teaching use in many ways, of which
error types, be taken into consideration, but a few only have been illustrated. Yet there is a
the forecaster’s task would be extremely broader and possibly more powerful argument
complicated if they were. Study of the in favor of acquiring such knowledge, for the
mistakes themselves seems to be a short cut. evident possession of it awakens the pupils’
sympathies and reassures them that the
There is much to be said on the other side, teacher is ‘on their side’. Few learners are not
however. Learners’ mistakes must inevitably pleased at the accurate, if occasional, use by
mean more to those who know something of the teacher of their own tongue; and a little
the learners’ language than those who are goes a long way. It suggests at least that he
ignorant of it. Using mistakes analyses, the is trying to see things from their viewpoint. It
latter have a good close-in view of what may makes clear that the learners’ language is not
go wrong, but can only guess at the looked down upon or regarded as irrelevant
underlying causes. Unless these are nuisance.
understood, the teacher is less aware of the
connection between one mistake and another, Both procedures are thus to be recommended,
and is thus less well equipped for the a study of the first language and also of
systematic treatment. Without some characteristic mistakes in learning the second.
knowledge of the learners’ language, Advanced knowledge of a language is not
moreover, it is sometimes hard to see what is necessarily accompanied, of course, by the
meant. Take, for instance, “They have arrived ability to analyze it. Unfortunately there are
there are three days”, a sentence that would many English-speaking teachers who assume
never be perpetrated by anyone properly that because they speak English they are well
taught. Could it mean “They have arrived and qualified to tech it: their ideas on the phonetic
are staying three days”? nobody would think and grammatical makeup of English are
85
sometimes extremely naïve. Non-English- with the question of learners’ errors and their
speaking teachers may have some similarly correction. It almost seems as if they are
crude notions of their own language. They in dismissed as a matter of no particular
particular, since most of them are importance, as possible annoying, distracting,
permanently occupied with English teaching in but inevitable by-products of the process of
one language area, should make as careful an learning a language about which the teacher
analytical study of that language as of the one should make as little fuss as possible. It is of
they teach. Otherwise the comparisons they course true that the application of linguistic
try to draw between the two languages will be and psychological theory to the study of
hesitant and often invalid. language learning added a new dimension to
the discussion of errors; people now believed
What cooperation can there be between the they had a principled means for accounting for
‘local’ teacher, possibly born and brought up these errors, namely that they were the result
among the kind of pupils he teaches, and the of interference in the learning of a second
teacher among and English-speaking country, language from the habits of the first language.
born and brought up among those who speak The major contribution of the linguist to
English alone? Each has his strong suit –the language teaching was seen as an intensive
former an expert acquaintance, the latter, a contrastive study of the systems of the second
special knowledge of the second. The most language and the mother tongue of the
obvious form of mutual assistance they can learner; out of this would come an inventory
give is to improve each other’s grasp of their of the areas of difficulty which the learner
own language. If both have had a measure of would encounter and the value of this
linguistic training, they can usefully cooperate inventory would be to direct the teacher’s
also in the collection and analysis of learners’ attention to these areas so that he might
mistakes. By referring the learners’ speech or devote special care and emphasis in his
writing to his own ideas on acceptable English, teaching to the overcoming, or even avoiding,
the one discovers and describes these of these predicted difficulties. Teachers have
mistakes. The other looks at the mistakes not always been very impressed by this
and, referring them to his experience of the contribution from the linguist for the reason
learners’ home language, describes them, if that their practical experience has usually
possible, in terms of failure to resist the already shown them here these difficulties lie
attraction of its phonetic, lexical, syntactic, or and they have not felt that the contribution of
other forms. Every mistake is thus seem from the linguist has provided them with any
two angles, and the resulting account of major significantly new information. They noted for
types of mistake will show clearly, in the example that many of the errors with which
context of any language area, what is most they were familiar were not predicted by the
difficult in English for the learners and what in linguist anyway. The teacher has been on the
the home language for the teachers. An whole, therefore, more concerned with how to
English-speaking teacher can himself, of deal with these areas of difficulty than with
course, excavate in the home language to the simple identification of them,, and here
uncover the causes of mistakes; but the ‘local’ has reasonably felt that the linguist has had
teacher may be able adequately to list and little to say to him.
describe the mistakes themselves; but on the
whole this appears to be the Englishman’s job. In the field of methodology there have been
Overlapping of the two activities is two schools of thought in respect of learners’
nevertheless essential if both parties are to errors. Firstly the school which maintains that
get the maximum benefit from this work. if we were to achieve a perfect teaching
Such a project for joint research is likely to method the errors would never be committed
have a practical outcome and to be worth in the first place, and therefore the occurrence
doing anywhere, but especially where the of errors is merely a sign of the present
rivalry between ‘local’ and ‘English’ teachers is inadequacy of our teaching techniques. The
more obvious than the cooperation. philosophy of the second school is that we live
in an imperfect world and consequently errors
The Significance of Learners’ Errors will always occur in spite of our best efforts.
Our ingenuity should be concentrated on
S. P. Corder techniques for dealing with errors after they
have occurred.
When one studies the standard works on the
teaching of modern languages it comes as a Both these points of view are compatible with
surprise to find out cursorily the authors deal the same theoretical standpoint about
86
language and language learning, and new hypotheses being set up to account
psychologically behaviorist and linguistically for the process of child language acquisition, it
taxonomic. Their application to language would seem reasonable to see how far they
teaching is known as the audiolingual or might also apply to the learning of a second
fundamental skills method. language.

Both linguistics and psychology are in a state Within this new context the study of errors
at the present time of what Chomsky has takes on a new importance and will I believe
called “flux and agitation” (Chomsky 1966). contribute to a verification or rejection of the
What seemed to be well-established doctrine new hypothesis.
a few years ago is now the subject of
extensive debate. The consequence of this for This hypothesis states that a human infant is
language teaching is likely to be far reaching born with an innate predisposition to acquire
and we are perhaps only now beginning to language; that he must be exposed to
feel its effects. One effect has been perhaps to language for the acquisition process to star;
shift the emphasis away from a preoccupation that he possesses an internal mechanism of
with teaching toward a study of learning. In unknown nature which enables him from the
the first instance this has shown itself as a limited data available to him to construct a
renewed attack upon the problem of the grammar of a particular language. How he
acquisition of the mother tongue. This has does this is largely unknown and is the field of
inevitably led to a consideration of the intensive study at the present time by
question whether there are many parallels linguists and psychologists. Miller (1964) has
between the processes of acquiring the pointed out that if we wished to create an
mother tongue and the learning of a second automaton to replicate a child’s performance,
language. The usefulness of the distinction the order in which it tested various aspects of
between acquisition and learning has been the grammar could only be decided after
emphasized by Lambert (1966) and the careful analysis of the successive stages of
possibility that the latter may benefit from a language acquisition by human children. the
study of the former has been suggested by first steps therefore in such a study are seen
Carrol (1966). to be a longitudinal description of a child’s
language throughout the course of its
The differences between the two are obvious development. From such a description it is
but not for that reason easy to explain: that eventually hoped to develop a picture of the
the learning of the mother tongue is procedures adopted by the child to acquire
inevitable, whereas, alas, we all know that language (McNeill 1966).
there is no such inevitability about the
learning of a second language; that the The application of this hypothesis to second
learning of the mother tongue is part of the language learning is not new and is essentially
whole maturational process of the child, whilst that proposed fifty years ago by H. E. Palmer
learning a second language normally begins (1917). Palmer maintained that we were all
only after the maturational process is largely endowed by nature with the capacity for
complete; that the infant starts with no overt assimilating language and that this capacity
language behavior, while in the case of the remained available to us in a latent state after
second language learner such behavior, of acquisition of a primary language. The adult
course, exists; that the motivation (if we can was seen as capable as the child of acquiring
properly use the term in the context) for a foreign language. Recent work (Lenneberg
learning a first language is quite different from 1966) suggests that the child who fails for any
that for learning a second language. reason, i.e., deafness, to acquire a primary
language before the age of twelve thereafter
On examination it becomes clear that these rapidly loses the capacity to acquire language
obvious differences imply nothing about the behavior at all. This finding does not of course
process that take place in the learning of first carry with it the implication that the language
and second language. Indeed the most learning capacity to those who have
widespread hypothesis about how languages successfully learned a primary language also
are learned, which I have called behaviorist, is atrophies in the same way. It still remains to
assumed to apply in both circumstances. be shown that the process of learning a
These hypotheses are well enough known not second language is of a fundamentally
to require detailing here, and so are the different nature from the process of primary
objections to them. If then these hypotheses acquisition.
about language learning are being questioned
87
If we postulate the same mechanism, then we the important evidence. As Brown and Fraser
may also postulate that the procedures or (1964) point out, the best evidence that a
strategies adopted by the learner of the child possesses construction rules is the
second language are fundamentally the same. occurrence of systematic errors, since, when
The principal feature that then differentiates the child speaks correctly, it is quite possible
the two operations is the presence or absence that he is only repeating something that he
of motivation. has heard. Since we do not know what the
total input has been we cannot rule out this
If the acquisition of the first language is a possibility. It is by reducing the language to a
fulfillment of the predisposition to develop simpler system than it is that the child reveals
language behavior, then the learning of the his tendency to induce rules.
second language involves the replacement of
the predisposition of the infant by some other In the case of the second language learner it
force. What this consists of is in the context of might be supposed that we do have some
this paper irrelevant. knowledge of what the input has been, since
this is largely within the control of the
Let us say therefore that, given motivation, it teacher. Nevertheless it would be wise to
is inevitable that a human being will learn a introduce a qualification here about the
second language if he is exposed to the control of input (which is of course what we
language data. Study of language aptitude call the syllabus). The simple fact of
does in some measure support such a view presenting a certain linguistic form to a
since motivation and intelligence appear to be learner in the classroom does not necessarily
the two principal factors which correlate qualify it for the status of input, for the reason
significantly with achievement in a second that input is ‘what goes in’ not what is
language. available for going in, and we may reasonably
suppose that it is the learner who controls this
I propose therefore as a working hypothesis input, or more properly his intake. This may
that some at least of the strategies adopted well be determined by the characteristics of
by the learner of a second language are his language acquisition mechanism and not
substantially the same as those by which a by those of the syllabus. After all, in the
first language is acquired. Such a proposal mother tongue learning situation the data
does not imply that the course or sequence of available as input is relatively vast, but it is
learning is the same in both cases. the child who selects what shall be the input.

We can now return to the consideration of Ferguson (1966) has recently made the point
errors made by learners. When a two-year-old that our syllabuses have been based at best
child produces an utterance such as “This upon impressionistic judgments and vaguely
mummy chair” we do not normally call this conceived theoretical principles where they
deviant, ill-informed, faulty, incorrect, or have had any considered foundations at all.
whatever. We do not regard it as an error in The suggestion that we should take more
any sense at all, but rather as a normal account of the learner’s needs in planning our
childlike communication which provides syllabuses is not new, but has not apparently
evidence of the state of his linguistic led to any investigations, perhaps because of
development at that moment. Our response to the methodological difficulties of determining
that behavior has certain of the characteristics what the learner’s needs might actually be.
of what would be called ‘correction’ in a Carroll (1955) made such a proposal when he
classroom situation. Adults have a very strong suggested it might be worth creating a
tendency to repeat and expand the child’s problem-solving situation for the learner in
utterance in an adult version; something like which he must find, by enquiring either of the
“Yes, dear, that’s mummy’s char”. teacher or a dictionary appropriate verbal
responses for solving the problem. He pointed
No one expects a child learning his mother out that such a hypothesis contained certain
tongue to produce from the earliest stages features of what was believed to occur in the
only forms which in adult terms are correct or process of language acquisition by the child.
non-deviant. We interpret his ‘incorrect’
utterances as being evidence that he is in the A similar proposal actually leading to an
process of acquiring language and indeed, for experiment was made by Mager but not in
those who attempt to describe his knowledge connection with language teaching (Mager
of the language at any point in its 1961, pp. 405-13); it is nevertheless worth
development, it is the ‘errors’ which provide quoting his own words.
88
 distinction between those errors which are the
“Whatever sequencing criterion is used it is product of such chance circumstances and
one which the user calls a ‘logical’ sequence. those which reveal his underlying knowledge
But although there are several schemes by of the language to date, or, as we may call it,
which sequencing can be accomplished and, his transitional competence. The errors of
although it is generally agreed that an performance will characteristically be
effective sequence is one which is meaningful unsystematic and the errors of competence,
to the learner, the information sequence to be systematic. It will be useful therefore
assimilated by the learner is traditionally hereafter to refer to errors of performance as
dictated entirely by the instructor. We mistakes, reserving the term error to refer to
generally fail to consult the learner in the the systematic errors of the learner from
matter except to ask him to maximize the which we are able to reconstruct his
effectiveness of whatever sequence we have knowledge of the language to date, i.e., his
already decided upon”. transitional competence.

He points out as the conclusions he draws Mistakes are of no significance to the process
from his small scale experiment that the next of language learning. However the problem of
step would be to determine whether the determining what is a learner’s mistake and
learner-generated sequence, or, as we might what a learner’s error is one of some difficulty
call it, his built-in syllabus, is in some way and involves a much more sophisticated study
more efficient than the instructor-generated and analysis of errors than is usually accorded
sequence. The problem is to determine them.
whether there exists such a built-in syllabus
and to describe it. In such an investigation A learner’s errors, then, provide evidence of
that the study of learners’ errors would the system of the language that he is using
assume the role it already plays in the study (i.e., has learned) at a particular point in the
of child language acquisition, since, as has course (and it must be repeated that he is
been pointed out, the key concept in both using some system, although it is not yet the
cases is that the learner is using a definite right system). They are significant in three
system of language at every point in his different ways. First to the teacher, in that
development, although it is not the adult they tell him, if he undertakes a systematic
system in the one case, nor that of the second analysis, how far towards the goal the learner
language in the other. The learners’ errors are has progressed and, consequently, what
evidence of this system and are themselves remains for him to learn. Second, they
systematic. provide to the researcher evidence of how
language is learned or acquired, what
The use of the term systematic in this context strategies or procedures the learner is
implies, of course, that there may be errors employing in his discovery of the language.
which are random, or, more properly, the Thirdly (and in a sense this is their most
systematic nature of which cannot be readily important aspect) they are indispensable to
discerned. The opposition between systematic the learner himself, because we can regard
and nonsystematic errors is important. We are the making of errors as a device the learner
all aware that in normal adult speech in our uses in order to learn. It is a way the learner
native language we are continually committing has of testing his hypotheses about the nature
errors of one sort or another. These, as we of the language he is learning. The making of
have been so often reminded recently, are errors then is a strategy employed by children
due to memory lapses, physical states such as acquiring their mother tongue and by those
tiredness, and psychological conditions such learning a second language.
as strong emotion. These are adventitious
artifacts of linguistic performance and do not Although the following dialogue was recorded
reflect a defect in our knowledge of our own during the study of child language acquisition
language. We are normally immediately aware (Van Buren 1967) it bears unmistakable
of them when they occur and can correct similarities to dialogues which are a daily
them with more or less complete assurance. It experience in the second language teaching
would be quite unreasonable to expect the classroom:
learner of a second language not to exhibit
such slips of the tongue (or pen), since he is Mother: Did Billy have his egg cut up for him
subject to similar external and internal at breakfast?
conditions when performing in his first or Child: Yes, I showeds him.
second language. We must therefore make a Mother: You what?
89
Child: I showed him. necessary at this point to posit a distinction
Mother: You showed him? between the two. Whilst one may suppose
Child: I seed him. that the first language learner has an
Mother: Ah, you saw him. unlimited number of hypotheses about the
Child: Yes, I saw him. nature of the language he is learning which
must be tested (although strong reasons have
Here the child, within a short exchange been put forward for doubting this) we may
appears to have tested three hypotheses: one certainly take it that the task of the second
relating to the concord of subject and verb in language learner is a simpler one: that the
a past tense, another about the meaning of only hypotheses he needs to test are: “Are
show and see and a third about the form of the systems of the new language the same or
the irregular past tense of see. It only remains different from those of the language I know’?”
to be pointed out that if the child had “And if different, what is their nature?”
answered I saw him immediately, we would Evidence for this is that a large number, but
have no means of knowing whether he had by no means all, of his errors, are related to
merely repeated a model sentence or had the systems of his mother tongue. These are
already learned the three rules just ascribed to interference from the habits of the
mentioned. Only a longitudinal study of the mother tongue, as it is sometimes expressed.
child’s development could answer such a In the light of the new hypotheses they are
question. It is also interesting to observe the best not regarded as the persistence of old
techniques used by the mother to ‘correct’ the habits, but rather as signs that the learner is
child. Only in the case of one error did she investigating the systems of the new
provide the correct form herself: you saw him. language. Saporta (1966) makes his point
In both the other cases, it was sufficient for clear.
her to query the child’s utterance in such a
form as: you what? Or you showed him? “The internal structure of the (language
Simple provision of the correct form may not acquisition) device, i.e., the learner, has gone
always be the only, or indeed the most relatively unexplored except to point out that
effective, form of correction since it bars the one of its components is a grammar of the
way to the learner testing alternative learner’s native language. It has generally
hypotheses. Making a learner try to discover been assumed that the effect of this
the right form could often be more instructive component has been inhibitory rather than
to both learner and teacher. This is the import facilitative” (p. 91).
of Carroll’s proposal already referred to.
It will be evident that the position taken here
We may note here that the utterance of a is that the learner’s possession of his native
correct form cannot be taken as proof that the language is facilitative and that errors are not
learner has learned the systems which would to be regarded as signs of inhibition, but
generate that form in a native speaker, since simply as evidence of his strategies of
he may be merely repeating a heard learning.
utterance, in which case we should class such
behavior, not as language, but in Spolky’s We have been reminded recently of Von
term (Spolky 1966) “language-like behavior”. Humboldt’s statement that we cannot really
now we must overlook the fact that an teach language, we can only create conditions
utterance which is superficially nondeviant is in which it will develop spontaneously in the
not evidence of a mastery of the language mind in its own way. We shall never improve
systems which would generate it in a native our ability to create such favorable conditions
speaker since such an utterance must be until we learn more about the way a learner
semantically related to the situational context. learns and what his built-in syllabus is. When
The learner who produced ‘I want to know the we do know this (and the learner’s errors will,
English’ might have been uttering an is systematically studied, tell us something
unexceptionable sentiment, but it is more about this) we may begin to be more critical
likely that he was expressing the wish to know of our cherished notions. We may be able to
the English language. Only the situational allow the learner’s innate strategies to dictate
context would show whether his utterance our practice and determine our syllabus; we
was an error or not. may learn to adapt ourselves to his needs
rather than impose upon him our
Although it has been suggested that the preconceptions of how he ought to learn, what
strategies of learning a first and second he ought to learn, and when he ought to learn
language may be the same, it is nevertheless it.
90
 2. There are several words that can take
Contrasting Spanish and English plural forms in Spanish where English has
Grammar Systems a mass noun:

INFINITIVES

1. Spanish often uses the infinitive as an

abstract noun which corresponds to the
English –ing form.
3. There are no irregular plurals in Spanish:

As a consequence, learners may have

difficulty in understanding sentences
which have –ing forms as subjects.
ADJECTIVES
2. Spanish has an infinitive marker (a), but
its distribution does not correspond fully to 1. Adjectives can stand without either a noun
English to: or a proform one:

ARTICLES 2. Comparatives and superlatives are always

formed with the equivalents of more and
1. the definite article goes with mass nouns most:
and plural count nouns that are used with
a general meaning:

RELATIVE PRONOUNS AND CLAUSES

2. Spanish uses the definite article with
possessive pronouns: 1. Spanish does not distinguish between
personal and non-personal relative
pronouns:

3. In some expressions singular count nouns

need no article:

2. The relative pronoun can never be deleted

in Spanish and learners have difficulty in
understanding English clauses with zero
4. Spanish makes no distinction between the relative pronoun:
definite article and the number one:

QUESTIONS
5. The definite article has a plural form.
1. There is not set word order for questions,
and auxiliaries play no part:

NUMBER

1. In Spanish the form of the plural is similar

to English (-s ending), but the ending is
used with articles, adjectives and
possessives as well as nouns:

91
2. Learners have difficulty with do/does/did? PERSONAL PRONOUNS

1. Subject personal pronouns are largely

unnecessary in Spanish because the verb
ending indicates person and number:

VERBS

1. Spanish has only one category of verbs,

and they all show the normal range of 2. In Spanish most personal pronouns have
tenses (present, past, future, conditional) the same form for subject and object:
and composite forms (progressive,
perfect). There is thus no separate
category of modal auxiliaries as in English,
and learners find the concept, the
simplicity of their forms, and their uses 3. There is no equivalent for the structure ‘it
difficult to grasp. Typical mistakes include: is + pronoun’ as used to identify oneself:

WORD ORDER

Word order is much freer in Spanish than in

English

2. In many cases where in English there are 1. ‘Subject-verb’ and ‘verb-subject’ do not
two different words, or a word is used with regularly correspond to statement and
two different structures, Spanish has question respectively:
verbs with both transitive and intransitive
possibilities: (got up, wait, do)

2. The freer word order allows words that are

emphasized to be placed last:

3. Particular problems of verbal structure

include the fact that many English phrases
consisting of ‘be + adjective’ are
expressed in Spanish by ‘have + noun’ 3. Frequency adverbs have several possible
positions, but not the typical mid-position
of English:

In Spanish ‘I like this’ is expressed as This

please me, leading to confusions such as: 4. Adjectives and noun typically postmodify
head nouns:

4. The Spanish endings corresponding to –

ing and –ed (adjectives) do not have
active and passive meaning respectively, 5. In Spanish an indirect object must have a
with consequent confusion in English. preposition (a) and the two objects can go
in either orders:

92
6. Adverbials and object complements are
regularly placed before a direct object:

NEGATIVES

1. Auxiliaries play no part in forming

negative sentences in Spanish, the
negative word goes before the verb
phrase:

The negative participle is no:

In short answers, the negative goes after

the pronoun, adjective, adverb, etc.:

2. The double negative is standard:

POSSESSIVES

1. Spanish uses definite article in some

sentences, not a possessive:

2. Spanish has the same word for yours

(formal), his, her, and their:

93
 A COURSE IN PHONOLOGY
Roca and Johnson

94
1. BILABIAL. Both lips they try to express themselves in English,
2. LABIODENTAL. Lower lip and upper teeth. especially during the initial stages of language
3. DENTAL. Tongue and upper teeth. acquisition.
4. ALVEORLAR. Tongue and gum ridge.
5. PALATAL. Tongue and hard palate. There are hundreds of textbooks for college
6. VELAR. Tongue and soft palate. classes in Linguistics that present a
7. NASAL. Velum lowered, nasal cavity ‘comparative’ approach to English and Spanish
opened. pronunciation patterns and other grammatical
8. GLOTTAL.. passage between vocal cords features of both languages. Professors
closed. (Donald) Bowen and (Robert) Stockwell, both
from UCLA, were among the first
http://www.bilingualquestions.org/qa207.htm ‘comparative’ linguists to explore the
similarities and differences between English
207. subject: phonological difficulties. My and Spanish. In many Spanish grammar
Spanish students have many difficulties with books –and in many English dictionaries- you
their phonetics. They use the phonemic can find detailed information about the sound
system of their native language when they system of each language.
speak English. I would like to do a
phonological analysis: find specific examples Many textbooks for college courses in English
of pronunciation mistakes that I have and/or Spanish Linguistics would help. You
observed in my students. How can you help may wish to contact your local college library
me to understand the differences between the and see what they offer.
two languages? Any particular book you could
recommend? THANK YOU for your interest in providing
the best instructional program for All
ANSWER your students!!!

YES!!! The native language background of any http://www.nichd.nih.gov/RFA/hd-99-012/hd-

English Language Learner is a very significant 99-012.htm
factor in the development of English language
skills. ALL English language learners use the
sound system of their native languages as
95
Comparisons of Oral Language and Literacy reading at various points in Spanish
Development (Reading and Writing) in reading instruction (e.g., by February and
Spanish and English: June of kindergarten, first, second, third,
and fourth grades)? Does this differ for
There is a compelling need to identify the children taught in Spanish only versus
ways in which the development of a first Spanish and English simultaneously?
language (e.g., Spanish) supports or • Are there differences in the rate and
interfaces with learning to read and write in a quality of oral language development in
second language (e.g., English). Currently, Spanish and English in Spanish-speaking
there is scant converging information that children with differing degrees of
bears on the relationship between knowledge bilingualism and biliteracy? If so, are these
of the first language linguistic, orthographic differences related to the rate and quality
and literacy skills and the nature, timing, and of Spanish and English language reading
quality of acquiring linguistic, orthographic, and writing development?
and literacy skills in the child’s second • Do Spanish-literate children make a more
language. Directly put, how does the rapid transition to skilled English language
development of literacy skills in Spanish differ reading and writing if they possess explicit
from the development of literacy skills in knowledge about the differences between
English and how do such differences inform Spanish and English orthographies?
instructional decisions? Studies that • Prior to instruction to develop English-
investigate the relationships between other language reading and writing skills, what
native oral and written languages (French, specific decoding, word identification, and
Turkish, Arabic, Japanese, Chinese, etc.) and comprehension approaches and strategies
English-language literacy development could do Spanish-literate students employ when
be highly informative vis-à-vis transition asked to read and write in English? Does
issues and are also encouraged if specific this vary as a function of age and grade
linkage can be made to the Spanish to English level?
transition. • What conscious metacognitive knowledge
about the relationships between Spanish
Longitudinal and cross-sectional studies that and English oral language and literacy
could address this priority are reflected in the concepts do students have when acquiring
following questions and issues. These English and Spanish literacy concepts
examples, are illustrative, not inclusive. simultaneously in contrast to when English
• Are there similarities and differences literacy instruction is provided following
between Spanish and English in the varying degrees of literacy development in
relative influence of specific oral-language Spanish? Is there a relationship between a
and orthographic skills (phonology, conscious understanding of English /
morphology, morpho-phonology, syntax, Spanish contrasts and the rate and quality
semantics) on reading and writing skill of English language literacy skill
acquisition at different phases of reading development?
development? • What proportion of variance in the rate of
• Are there differences between Spanish and English-literacy acquisition can be
English in the degree to which different attributed to the level of development in
language and print concepts predict Spanish-literacy skills, Spanish oral
reading and writing developmental levels language skills, and English oral language
in the native language? If so, what are the skills? Is the level of Spanish literacy
implications for instructional decisions in and/or oral language skills at the
both Spanish and English? introduction of English literacy instruction
• Does the development of Spanish oral- a better predictor of English literacy
language skills vary as a function of, for acquisition than the rate of ongoing
example, geographic region, socio- acquisition of Spanish literacy skills, or
economic status, generational status, and vice versa?
cultural development of English oral
language?
• How mush knowledge and what aspects of
oral language and reading have students,
on average, acquired about Spanish

96