Está en la página 1de 25

10 Malinowski and Firth

10.00 Bloomfield and Sapir were agreed that linguistic study of Speech
uncovers the form of Language. In quite different senses, both could be said
to consider that study of language reveals a determinant of the form of mind.
Sapir devoted comparatively more attention to the interdependence of
Language and Culture. More than Sapir, Bloomfield found linguistic form
definable in positive physical terms, negatively in behavioral or psychological
ones. Since mental structures were as physical for Bloomfield as things
outside, he faulted Sapir's conceptual account of speakers' 'insides'. Those
insides cannot be autonomous. They simply must be subject to the same laws
as those outside a head, though more difficult to observe. Yet his brief review
of Sapir's Language conceded that there was no other way of talking about
such things to a general audience in 1921 (see Hockett 1965.91ff).
So if it is only a plausible metaphor to say there is an (a) science like
Mentalist Psychology to study that particular part of the world which is located
at (A), there is no real harm in calling Behaviorism a separate science (B),
and Linguistics yet another science (C). The subterfuge could conceal from
us that real Science must be Unified Science—perhaps worth calling Science
(D)-with (a) (b) (c) as subsets. For the time being, linguistics might use
psychological, inside, methods and terms, which are reducible in Science (B)
anyway, and that is why Bloomfield could say we 'must act as though' all
findings from (A) and (B) are in, or imminent.
The linguistics of J. R. Firth is an attempt to study (a), (b), and (c) more
as a unified, interdependent complex, rather than as independent elements
related in unspecified ways. That makes linguistics a single discipline, the
object of which involves the triad (a)-(b)-(c). Another conception of that
totality, trying to do professional justice to its hitherto distinguished aspects,
is found in concepts Bronislaw Malinowski developed for anthropology.

10.01 Bronislaw Malinowski (1884-1942). Malinowski was a facile


polyglot rather than a linguist. He early acquired a knack for languages from
his father (a pioneer in Polish language and dialects) and from his multilingual
mother (in Mediterranean travel after his father's early death). He first
studied physical sciences, but read omnivorously. To master English, he read
Frazer's 1890 Golden Bough, and was intrigued, as his publications show, by
its account of the evolution of magical, through religious, to scientific thought,
from primitive to modern societies:
314 / GENERAL LINGUISTICS
For no sooner had I begun to read this great work, than I became immersed in it and
enslaved by it. I realized then that anthropology, as presented by Sir James Frazer, is a
great science, worthy of as much devotion as any of her elder and more exact sister-
sciences, and I became bound to the service of Frazerian anthropology. (Malinowski
1948.72)

Malinowski won his Ph.D. in 1908 with the highest honors in the then
Austrian Empire. After two years of postdoctoral studies at Leipsig
(economics with Bucher, psychology with Wundt), he enrolled in the new
program in Anthropology at the London School of Economics, and was
awarded its doctorate in 1916 for his 1913 Family Among the Australian
Aborigines. Meanwhile, he published extensively in English and Polish,
notably book reviews on topics anthropologically alien, but later incorporated
into his own theories. These publications made of him a sort of
interdisciplinary ambassador for Anthropology. His Trobriand Island research
(1915-16, 1917-18) gave him international stature, materials for years of
publications, and data for much of his theoretical work.
He seems rarely to have made a neutral impression. People were
attracted or repelled in about equal force and numbers, in private or
professional life. For example, though technically an enemy alien during
World War I, the Australian Government helped finance his Trobriand
research. He published constantly, is reckoned a founder of Social
Anthropology, popularized it worldwide as the functional study of culture—
real people satisfying vital needs. In the process, he was the teacher of most
British and Commonwealth workers in that field between the two World
Wars. His theoretical standing is debated.

10.02 Eclecticism. Malinowski adapted insights from diverse scholars to


form his original synthesis. Influences he explicitly mentioned include Frazer
(but not his determinist, evolutionary, ideas), Durkheim (while rejecting his
abstractionism), Freud (though he much modified the pansexual), and Hull
(but without the mechanist side of his Behaviorism). What he drew from
them were holistic, internal-defining, as opposed to external-imposed,
structural conceptions, such as those de Saussure expressed so well for
language. But abstractions like de Saussure's langue or Durkheim's collective
conscience were counter to his professed concern for empirical data.
Critics see his terminological shifts as evidence either of inconsistency or
an ability to adapt to new evidence. He modified earlier assumptions about
cultural superiority. His early titles had 'savage' or 'primitive' labels. His
views about societies in the 1923 Ogden and Richards Supplement are not
those of his 1935 Coral Gardens. His Encyclopedia Britannica article on
Anthropology (13th ed., suppl., 1926.133), stressed functionalist assumptions:

...in every type of civilization, every custom, every material object, idea and belief fulfills some
vital function, has some task to accomplish, represents an indispensable part within a working
whole.
10 Malinowski and Firth /315

His posthumous Scientific Theory of Culture (1944) laid out relations


among things and thought. Needs for individual survival (like food or air)
differ from those required in group behavior. Values set by the group are
embodied in symbols that demand, prohibit, or tolerate different responses,
facilitating social evaluation. These needs, in his view, are:

...the system of conditions in the human organism, in the cultural setting, and in the relation
of both to the natural environment, which are sufficient and necessary for the survival of group
and organism. A need, therefore, is the limiting set of facts. Habits and their motivations, the
learned responses and the foundations of organization, must be so arranged as to allow the
basic needs to be satisfied. (1944.90)

Sapir pointed out the function of language as a substitute for symbolic


action, Malinowski stressed biological determinants of cultural activities, and
the delicate hierarchy of primary, derived, and integrative needs. When
primary needs are frustrated, cultural change is imminent. Optional
satisfaction of derived needs allows individual diversity, a tolerably similar
and cohesive world-view—even one embodying beliefs strange to us—is a
condition for group stability. The complexities with which he dealt are
reflected in divergent views of Malinowski's work in Man and Culture (R.
Firth 1957). This book evaluates Malinowski's personal and scientific
standing, his concepts of culture, law, kinship, magic and religion, economics,
social systems and change, needs, the conduct of field work, and analysis of
language. Those views are not as simple as (a), (b), (c).

10.3 Purpose and Culture. The defense of teleology in culture ('some


task to be accomplished') is not dominant in Malinowski's writings. But it was
just that theme in Darwin's evolutionary hypotheses that critics found a
backward step in natural science. It might have been explained by
distinguishing inanimate nature and social reality, or by suggesting that
society's purposes are no more or less inexplicable than its social contract.
Malinowski might have said it was just a manner of speaking, as Bloomfield's
review of Sapir's Language held. While psychologizing was not scientifically
respectable, it was the only way of talking about language generally
understood. ((1922), Hockett 1965. 91ff) In either case, it is worth recalling
our earlier discussion of the nature of explanation which empirical science
sought to replace, and weigh whether Malinowski's Functionalism was
retrograde or compatible with either.

10.4 Description and Explanation. Western thought long accepted one


basic conception of explanation from Aristotle: complete understanding of
something resulted when its four causes were known. The four were labelled
material, formal, final, and efficient causes. Cause is an analogical term in
these four collocations. Aside from substituting matter for the first, and
structure, pattern от function for the second, empirical science found little use
for the other two.
316 / GENERAL LINGUISTICS

A familiar notion of cause was efficient cause. Its clearest example was
deliberate human initiation of change, observable in common experience.
This efficient causality effects accidental changes by rearranging matter, and
its products are artifacts like a better mousetrap. Whatever is capable of such
(re)formation is then viewed as material cause, its structure or arrangement
as a formal cause, and the end or purpose of the artifact as its final cause.
When dealing empirically with actual material instead of ideal matter, and
with phenomena in nature rather than human products, physical sciences study
a different object than metaphysical 'sciences'. Until well into the modern
era, as suggested in Chapter 9, philosophic was the prestige expression and the
method most comparable to what would be called scientific today, particularly
when applied to explanation.
None of these causes is directly experienced in what we see, feel or hear.
Yet sense data is what all humans have in common. Their subjective mental
worlds may seem either irreconcilable or incommunicable. They are
acknowledged to be inaccessible externally in any account. Mental cause may
effect subjective conviction, but empirical causes are expected to produce
results anyone can observe.
It would be retrograde to look for scientific explanation outside the
system empirical science has established. So its legitimate data must be
sensible: evidence must be empirical; explanation is or involves predictive
success for observable cooccurence. From the empirical point of view,
traditional notions of cause are not objects for experiment or observation, but
for philosophic conjecture.
Human purposes and responsibility are extrasystemic considerations in
natural science. Their composition, distribution or function are not empirical
objects. Traditional discussion dwelt more on the composition of these
concepts, their distribution and function. As structurally determined and
mutually defining elements of intelligibility, they are very much intrasystemic
factors when description is held to be different from explanation.
So a curious tension arises when rival approaches to the same data, with
different goals, confront each other. Sometimes previously accepted causes
are declared pseudo-causes; or predecessors defending them are told they
have been merely describing, while successors will now explain.
This is quite like the atmosphere of Jakobson's teleological phonology
confrontation with phoneticians at the First International Conference of
Linguists (1929): Phoneticians were informed that they had been doing the -
etic donkey work of description, and the -emic work of phonological
explanation in terms of function could now take over. Similar 'demotions' are
exemplified in the suggestion that syntax could supply the missing explanatory
element in largely morphological, 'mere' descriptive work; proposals that a
semantic perspective can illuminate 'mere' syntactic description; or that
pragmatics can make 'mere' abstract cognitive semantics more realistic.
10 Malinowski and Firth /317

10.05 Explaining Language Use. Malinowski's functional anthropology


claimed to explain what others just described—poorly—because they either
ignored teleology, or imposed putative rational goals of Western society on a
(perhaps) irrational group. His work had the generic structural stamp of
working within observable data. He professed not to assume the goal is
known, and to refrain from judging societies according to their approximation
to it. Much of his theory about language was consistent with that orientation.

So his reformulations of what to look for in language use, and how to


regard it, were novel and refreshing. Some of his technical terms were often
felicitous in making neglected aspects of language use obvious. Such are his
use of context, meaning as function, and particularly one term he coined,
photic communion, discussed in the status of sentence and word.

10.06 Linguistic Fact. His approach to language is summed up in this


single passage on what he saw as central problems in translation:

It might seem that the simplest task in any linguistic enquiry would be the translation
of individual terms. In reality the problem of defining the meaning of a single word and of
proceeding correctly in the translation of terms is as difficult as any which will face us. It
is, moreover, in methodological order not the first to be tackled. It will be obvious to
anyone who has so far followed my argument that isolated words are in fact only linguistic
figments, the products of an advanced linguistic analysis. The sentence is at times a self-
contained linguistic unit, but not even a sentence can be regarded as a full linguistic datum.
To us, the real linguistic fact is the full utterance within its context of situation.
(Malinowski 1935: 1965:11)

10.07 Contextual analysis. Society makes it easier for humans to


function by integrating individuals into a group. Functions are satisfactions of
human needs, some universal, some individual. Nonfunctional are objects,
practices or events which satisfy no need. Language is a functional component
of society, and 'meaningfuP to the extent that it does something about needs.
Primitives, Malinowski had said (1915), have little or no need for philosophic
reflection. So that cannot be an important 'meaning' of language: it must be
something much more pragmatic. Survivors have an elemental need to
survive. They don't think much, Malinowski suggested, and if they do, they
don't talk about it.
An elementary need in society is cooperation and avoidance of discord.
Since he found people react more to attitudinal than intellectual factors, he
concluded that photic communion would be an apt name for that language use
which primarily cultivates unity, rather than imparting information. The words
in Have a nice day! or Love you! might elsewhere concern facts about the
world. But both expressions function commonly as just the right noises to
make, given the occasion and the need for relations, casual or binding. They
do not disclose genuine information. Their primary function is social.
318 / GENERAL LINGUISTICS

Given that assumption, the word, and less obviously, some sentences, are
'products of sophisticated linguistic analysis'. The real linguistic datum is the
sentence (more or less traditionally defined) in its context of situation, with
that situation in its context of culture as its immediate physical environment.
Linguistic elements are distinguishable to the extent they function—have some
observable effect—in the overall context of cultural process. Language use is
one form of cultural functioning.
So translation cannot be just matching words comparable in reference, or
related in sense, since both differ even among the cultural siblings of Europe:
Malinowski cites as imcomparables, English honor, German Ehre, French
honneur, and Spanish honra\ as untranslatable, English gentleman от fair-play,
German Weltschmerz, Italian bel canto and French connoisseur. (1935,1965.12)
De Saussure's distinction of signification from value and content implied just
that, as did Boas' analysis of The man is sick. Sentence translation based
only on grammatical approximation is insufficient: it would further obscure
value vs. content differences. Of course, if translation is paraphrase and
extensive commentary—his own practice—then it is possible. It may be better
than the word-for-word procedure, but it is still only an approximation.
Readers still register that information, given in their own language, against the
background of a very different culture.

10.08 Abstraction. These observations remind us that the problem for


any science is how abstract it can afford to be yet still remain intelligible, or
how concrete it can afford to get, without getting lost in details. Malinowski:

But it is easy to become redundant in commentaries and by no means obvious where


to draw the line between giving too much detail on the one hand and giving an insufficient
and altogether dry indication to the reader. (1935.1965.11)

His approach is intelligible, but it combines (a), (b), and (c) data without
an autonomous (or even subsidiary) science (C), about whose nature or
possibility we are presently concerned. (A) and (B) scientists can best judge
their representation in that scheme. In more recent terms, it proposes a study
of performance without one of competence; Firth assumed Malinowski
rejected the comparable distinction of langue and parole:

...he explicitly dissociated himself quite early from Durkheim's philosophical basis of
sociology [1913b; 1916.423.П.1]. He would have nothing to do with a collective soul and
presumably had no interest in the French conception of langue as a function of the
collectrvite... he declared that the postulate of a collectivity was barren and absolutely
useless for an ethnographic observer. (R. Firth 1957.95)

Malinowski's formula for stating 'meaning' is identical to the Behaviorists'


(the sum of practical events preceding and following an utterance). But what
actually precedes includes the speaker's and hearer's history since conception,
what follows is the rest of the world's history. Restricting them verbally by
10 Malinowski and Firth /319

relevant and practical specifies neither the degree of abstraction required, nor
its basis, nor practical techniques for isolating them. Relevance within a field
defines it: phonology's vague notion of 'different in meaning is easier to
specify in speakers' overt behavior than in even briefly delayed cultural impact.

His stress on the function of utterances suggests that their entire system
has its ultimate cause—its teleology or purpose—outside an empirically
observable system. Or, 'meaning in this system is equated with factors
subsequent to utterance: certainly not words, nor even sentences can be
assigned meaning antecedent to utterance within the logic of his approach.
An autonomous science of language ought to be able to do something like
that. And since Malinowski troubles to write about these things, it would seem
he assumes 'meaning must have just the interpretation he rejects, or no one
could understand or disagree with him. His proposal is like the Sapir-Whorf
Hypothesis: if valid in an unqualified sense, it would either have been
impossible for Sapir or Whorf to suspect it, or for anyone else to grasp what
little they had understood of it.
What Malinowski does, then, is simply to define the goals, scope and
purposes of a study involving what interests him about language (in a
nontechnical sense) and incorporate that, legitimately enough, into his
ethnographic theory of Language. He described or referred to it often enough
as a theory of language use. He clearly hoped it would bring to light some
linguistic universale, and contribute to the understanding of human
communication in general. This again raises the question of whether there
can be a legitimate science of language which abstracts from its use, social or
intellectual. Malinowski made an unqualified claim:

Translation in the sense of defining a term by ethnographic analysis... is feasible and


is the only correct way of defining the linguistic and cultural character of a word. (1935,
1965.17)

10.09 Induction and deduction in Malinowski. In defending his


originality on the role of magic and religion Malinowski had encountered in
Frazer's Golden Bough, Symmons-Symonolewicz (1960: v.4.36-43) quotes from
his own translation of a 1915 book of Malinowski's on the subject. This
antedates both his Trobriand field work and Loisy^ book, from which
Radcliffe-Brown thought Malinowski had borrowed.
Given previous allusions to how rationalists and empiricists can view each
others' work, it is instructive to see that Malinowski intended this work to
refute the 'rationalistic approach to the religious phenomena' (1960.40):

Man, especially primitive man who lives in a constant struggle for survival, cannot be
and is not a reasonable and reasoning being... His life is mainly emotional and active, full
of emotions and passions, and it is these elements that shape his whole behavior, and not
a philosophical reflection... Primitive man has urgent strong needs, constant, sometimes
dangerous, vital pursuits, and it is easy to show that these very elements lead him to the
320 / GENERAL LINGUISTICS
performance of such acts and activities which constitute a germ of religion. Up to this time
religion appeared to be an artificial theory developed out of speculations of the primitive
man... To a modern student primitive religion is not any more an artificial theory... but
rather a biological necessity for the man, a result of his struggle for that existence. (1960.40-
41)

This raises the question of when and how insights occur. Leach (in R.
Firth 1957.119ff) contradicts this early Malinowski from his reading of later
texts, where he finds that It was dogma for Malinowski that all human beings
are reasonable (sensibly practical) individuals (127), and that in Malinowski's
attempt to impose 'rationality' upon his savages... (he found) that the
Trobriander was more rational than himself. (128)
Empiricism as Bloomfield presents it comes to sober inductive generaliz-
ations after all the objective facts have been so marshalled, that no other
interpretation of observed co-occurrences would make sense. If Rationalist
thought is exemplified in Traditional Western Grammar, to which other
linguistic facts were expected to conform, its civilized consistency might tempt
one to ignore, deny, or explain away any divergence as primitive approxim-
ation. But both points of departure seem complementary, when faced with
the need to adapt—or adopt—some theory to explain the facts they describe.

10.10 Objectivity. Training must equip scientists with some system of


initial categorization before experience of concrete data. For objectivity, that
method must be self-correcting, as Boas' fieldwork training of others insisted.
In a scientific report, we expect nothing that cannot be publicly replicated or
falsified. Since few could go to the Trobriands to verify Malinowski's reports,
it would seem reasonable to allow, even expect, him to mention what it was
that 'compels us' to correlate language and activity and what 'forces us to
define meaning in terms of experience and situation' (1935.1965.9), or when it
was that he began to see the need of modifying his preconceived ideas.
This 1935 account of a Trobriand experience illustrates Malinowski's flair
for dramatizing previously dull anthropological work:

...I realized then and there what the real function of magic is. On the psychological side
it leads to a mental integration, to that optimism and confidence in the face of danger
which has won to man many a battle with nature or with human foes. Socially, magic, by
giving leadership to one man, establishes organization at a time when organized and
effective action is of supreme importance. (1960.44, quoted from Malinowski 1935)

Raymond Firth's assessment of Malinowski's theory/fact posture suggests


a similar interpretation:

At that time, the tradition was that an anthropologist was primarily either a
theoretician or an ethnographer, and that the theory should be kept separate from the
facts. It was part of Malinowski's contribution, not only to combine them, but to show how
fact was meaningless without theory and how each could gain in significance by being
consciously brought into relation. (R. Firth 1957:2)
10 Malinowski and Firth /321

10.11 John Rupert Firth (1890-1960). Firth took a Master's degree in


History and Language at the University of Leeds in 1913, and was professor
of English in Punjab in 1920. At the University of London, he studied and
taught Phonetics under Daniel Jones for ten years (1928-38) after which he
joined the School of Oriental and African Studies. He became head of its
Department of Phonetics in 1941, and, awarded the first Chair of General
Linguistics in England in 1944, he and Jones were important in spreading
respect for linguistics as an academic discipline. He retired in 1956.
Jones, Sweet, and the empirical tradition of British scholarship were
among his formative influences. Others included personal contact with
Malinowski, the challenge of vindicating linguistics as a peer discipline with
standard University studies, his interests in the history of linguistics, and in its
autonomy as a science. The latter two help account for his anti-structuralist
stance, and the development of an approach to linguistics characteristic of the
London School. Aside from two popular books on general linguistic topics,
his publications were mainly technical papers. But he engaged in exchange of
ideas with other scholars and with his faculty, carefully monitoring their
publications on topics suggested by his reflections on linguistics.
Exemplification of his ideas on prosodic analysis, for instance, is found in
papers by colleagues rather than in his own. He seemed to communicate best
an air, an attitude, a slant, or emphasis, on the need for studying actual
language in actual use, and a concern for a holistic concept of 'meaning' as a
modality of all levels of analysis.

10.12 Meaning. Like Malinowski's, Firth's conception of the goal of


linguistics can appear romantic, impossible, self-defeating, but attractively
concrete. That assessment would be inevitable if the object of linguistic
analysis were an entire language. But Firth saw its data as appropriately
restricted languages, dialectal or stylistic subsets within languages, or in familiar
situations like preaching or military drill. Others included the situation of
requesting something at table: some pout, or poke, or point; others feign
terminal embarrassment and inquire lengthily about the compatibility of lofty
personages with the menial work of transferring objects. In much the way
that lower levels of analysis have been portrayed as subordinate to higher ones,
Firth took 'meaning' to be the holistic function of speech and its components:

It can be described as a serial contexualization of our facts, context within context,


each one being a function, an organ of a bigger context and all contexts finding a place in
what may be called the context of culture. (Technique of Semantics 54)

Firth prided himself on his own empiricism, and extolled that tradition in
British linguistic work. The highest accolade he awards in his evaluation of
Malinowski was to show that he was almost British in that regard. Like
Malinowski, Firth borrowed or absorbed much from the insights of others, but
dissociated himself from what he considered their unempirical and empirical
failings. Malinowski's material contexts, for instance, became abstract
322 / GENERAL LINGUISTICS

categorizations for Firth; his monism was not the 'foolish bogey' Firth found
in the antimentalism of Bloomfieldians, but was dictated on empirical grounds,
since everyone claimed to know so little about minds. (J.R.Firth 1957.7)
Unlike Malinowski, he saw the need for, and set about outlining, an
independent science of language. Its task was the statement of 'meaning'.
The levels of linguistic analysis others approached as external to, only
distinctive of, or the setting for, meanings (phonetics, phonology, lexicon,
morphology, syntax, situation), he proposed to regard as constructs whose
function it was to allow the dispersion of unitary meaning into convenient
factors for specialized study. His suggestive analogy was the prismatic
dispersion of light into the distinct colors of the spectrum: linguistic theory
and its categories are to solidary meaning as the prism is to light-both human
artifacts. Reversing perspectives guarantees 'renewal of connection' with
experience.

10.13 Context of situation. Firth notes that Malinowski credits Philip


Wegener (1885) as the first to propose an interesting version of a contextual
approach to meaning. This required neither postulation of Durkheim's
collective soul or Saussure's langue, only observation of factors individuals
appreciate as institutionalized in their own society, and reinterpretation by the
scientist. While Malinowski could focus on culture, the linguist must
concentrate on texts. The point of departure should be institutionalized
concepts found in native words and taxonomies rather than those of the
investigator. Hence the importance Malinowski attached to mastering the
native language, and the stress Firth laid on real rather than fanciful linguistic
data. He approved of Malinowski's interest in Temple's (1899) sketch of a
Universal Grammar. Malinowski borrowed the technique of interlinear glosses
with extended commentaries from Temple—perhaps the first and chief
'linguistic' formation he acquired by the reading he reported doing in
Melbourne, between his first and second visits to the Trobriands. Temple
said:
Of course, grammarians will know that all this is syntax, and I will now explain why
I consider that it is far more important to study function than form as essential to the
correct apprehension of words, and how to my mind accidence arises properly out of syntax
and not the other way round, as we have all been taught... I found myself, in building up
the theory, compelled, in order to work out the argument logically, to commence where the
accepted Grammars ended, viz. at the sentence, defining the sentence as the expression of
a complex meaning, and making that the unit of language. (1899a.2: in R. Firth 1957.98)

J.R. Firth found Malinowski made little use of those insights and
remained 'reasonably traditional, but grammatically unsystematic' (R. Firth
1957.98). He agreed with Malinowski that a functional method was comple-
mentary, rather than antithetic, to the then predominant historical approach:

... the development of descriptive linguistics on a large scale is an essential preliminary for
the reformulation of problems in comparative and historical work. This could only be the
10 Malinowski and Firth /323
case if...linguistics recognizes that its principle objective is the study of meaning in its own
terms. (R. Firth 1957.100-1)

That tidy distinctions between purely empirical or rational approaches are


to be taken with a grain of salt emerges from his conclusion: empirical
Malinowski contributed little to his Ogden/Richards Supplement goal:

We need a Theory, devised for the purpose of observation of linguistic fact. This
theory would give a recast of grammatical definitions, based on an analysis of meaning. It
would analyze the nature of syntax, parts of speech and the formation of words, and
besides giving adequate and plastic definitions would open up vistas of problems and thus
guide research... And here, I cannot refrain from repeating a favorite quotation from
Goethe: Das Hochste ware zu begreifen, das alles Faktische schon Theorie ist. [The most
important thing is to understand that "facts" are only "facts" in terms of a theory.'] (J.R.
Firth 1957.101)

10.14 Facts. Firth's quotation from Goethe more or less repeats what
Malinowski often insisted on, that facts are only facts in the light of some
theory. This insight so impressed Firth that he chose it as the heading for his
retrospective Synopsis of Linguistic Theory 1930-1955, published a year after
his retirement. (Studies in Linguistic Analysis 1957) Firth distinguishes his
situational approach from Malinowski's, Gardiner's, and Wegener's:
Malinowski's was enthnographic, his own, linguistic; Gardiner's position was
close to Wegener's, yet differed on the basis of its 'realism'. Wegener
discusses three things basic to a situation as determining what one says: (1)
what there is to see, (2) what one actively recalls, and (3) concomitant
awareness of one's whole state of mind. Firth's approach eliminated two
elements of Wegener's triad (present recall and consciousness of personal
identity) but retained 'the objective situation as presented and observed', since:

A serious confusion of the analysis of the context of situation with the other levels of
analysis such as the grammatical level, has been one of the main weaknesses of early
attempts to relate statements of meaning to other social and psychological factors. (J.R.
Firth 1957.103)

10.15 Ecology. What all these approaches have in common is


comparable to ecological concerns: everything interpenetrates and affects
everything else. This is the essential structuralist view. In The Tongues of
Men ((1937) 1964.34)), Firth used the term implications to stress solidarity
among parts of Language, illustrated by Saussure-like associative relations
(35), the formation of sets peculiar to personal expectations (90), Pareto's
residues and derivations (96 ff.), Carnap's logical syntax (105 f.), and
Whitehead's concept of mutual prehension (110f.). He finds other
contextualists 'did not grasp the full implications of Wegener's hints... A general
theory such as this must include similar approaches in other branches of
linguistic analysis', (in R.Firth 1957.103)
324 / GENERAL LINGUISTICS

10.16 Firthian contextual analysis vs. Malinowski's. Firth denied


Malinowski's techniques were linguistic analysis:
This expression as used by linguists refers to highly abstract analyses of a given
language—usually a restricted language—at the phonological level, at various grammatical
levels and in the summary entries of dictionaries. Malinowski fully realized his
shortcomings in linguistic analysis as we now understand it, and said so explicitly. The
analysis to which I now proceed can be given only in an approximate manner, for in a full
one, a long disquisition on grammar would have to be given first.' He never managed to
realize what may have been his secret ambition - a technique of analysis satisfying the
demands of linguistic science. (R. Firth 1957.105-6)

Malinowski's technique was to give (1) an interlinear translation that


matches native words and formatives with an English 'fixed equivalent', (2) a
free translation, (3) a combination of (1) and (2), then (4) a detailed
commentary called 'the contextual specification of meaning', with phonetic and
grammatical notes. Firth rejects fixed equivalents as he did the distinction of
primary and derived meanings (cf. Bloomfield 1933.149ff). Collocational
studies (e.g. to determine interpretations of English ass) show that word-
meanings are not fixed, but context-dependent in Structures; that words
derive their meanings from sentences, not vice versa; that they can be 'fixed'
to cover systems (as he defined structure and system) is another matter:

10.17 System and Structure.


Systems of units or terms, set up by the linguist, provide sets of interior relations by
means of which their values are mutually determined. In order to have validity, such
systems must be exhaustive and closed, so far as the particular state of the language,
suitably restricted, is under description. (R. Firth 1957.107)

...the terms structure and system (are) distinct in technical use. Structures are abstractions
from utterances or parts of utterances recorded textually. Thus CVCVC and Noun-Verb-
Noun might each constitute a structure specifically defined in a particular language at the
phonological and grammatical levels respectively. A structure is said to comprise elements
or categories in mutual syntagmatic relation. At any given level of analysis closed systems
of categories, units or terms are set up to give mutually determined values to the elements
of structure. The terms of a system, or of a sub-system within it, commute, thus enabling
account to be taken of the elements, constituents and features which are given order and
place in structure. (R. Firth 1957.107 f.n. 1)

Malinowski gave six Trobriand words 'equivalent' to English garden. Firth


would accept them as a lexical system. In combining free and literal
translations, Malinowski makes additions, or he coordinates and subordinates
sentences differently with English conjunctions which are absent or vaguely
represented in native texts (1922a.458; 1935 11.38; 1957.107), a 'double-entry
procedure' which Firth finds the same as what is 'nowadays described as the
"translation meaning".'
Malinowski's steps 3 and 4 concern text, not situation, and Firth denies
they are based on a definite grammatical and lexicographic scheme, since the
10 Malinowski and Firth /325

distinction between inclusive/exclusive plurals and duals noted elsewhere is


not observed. (107 on 1922a.440) Malinowski's 'contextual specification of
meaning' confuses phonetics, grammar, and lexicon, plus situational factors
needed to account for what he calls the 'telegraphic', context-dependent, nature
of the language. But for Firth:

Grammar is concerned with the interrelation of categories, not of the words as such,
and cannot be derived from any context other than that of grammatical analysis. (R. Firth
1957.109)

He finds Malinowski's grammar notional, not linguistic. Saying a phonetic


element b may connect a verb with future, or perhaps potentiality in any tense,
or perhaps just give emphasis, confounds levels of analysis and does not justify
the English gloss might. (Coral Gardens 11.31). It is more notional than
phonetic to say the sound l connotes definiteness, or 'places the action into a
regular past, accomplished state... at times only gives emphasis... The letter 11
have rendered by the fixed meaning "did"...' {Coral Gardens 11.32)

Malinowksi tacitly accepts grammatical categories as universals when he


says transitive and intransitive verbs are hard to distinguish, or that passive is
lacking. While he deals better with classificatory particles, he has no canonical
lexical entry form (108). Malinowski knew his phonetics was superficial, but
debated inconclusively with himself whether more phonetic detail might not
be counterproductive. He resorts instead to impressions: 'alliteration dear to...
thumping rhythm indicated by sharp and circumflex accents... perfunctory
performance... fewer melodic modulations and phonetic peculiarities...
phonetically expressive... vowels with Italian values and impressive ring'.

10.18 Malinowki's strong points. Firth lists Malinowski's positive


contributions to linguistics under four headings:

I. General theory, especially his use of the concepts of context of


situation and types of speech function. (Coral Gardens 11.53;
1923a.475-7)
II. The statement of the meaning of a word by definition with
reference to culture context.
III. The statement of meaning by translation.
IV. The relations of (i) language and culture; and (ii) linguistics and
anthropology. (110)

10.19 Constructs. Compared then to his own linguistic theory,


Wegener's arealism contrasts with Gardener's and Malinowski's realist
approach, and that is an oddity, given Malinowski's stress on theory. It
suggests that brute facts exist independent of and prior to any statement of
fact, as quoted above: 'to us, the real linguistic fact is the full utterance within
326 / GENERAL LINGUISTICS

its context of situation'. Firth has thought out what he considers a defensible
relation between theoretical constructs like Circle, and the data of experience
stated in terms of them, e.g. Tires are circular. Apparently Malinowski had
not, since Firth finds that

...the word utterance seems to have had an almost hypnotic suggestion of reality which
often misleads him into the dangerous confusion of a theoretical construct with items of
experience. The factors or elements of a situation, including the text, are abstractions from
experience and are not in any sense embedded in it, except perhaps in an applied scientific
sense, in renewal of connection with it. (Ill)

Firth says he first used context of situation in his Speech (1930); his more
recent position makes explicit that

... the context of situation and the notion of types of speech function are best used as
schematic constructs to be applied to language events and that they are merely a group of
related categories at a different level from grammatical categories but of the same abstract
nature. The linguist sets up interior relations of three kinds:

(1) the interior relations of elements of structure, words and other bits and pieces of the
text;
(2) the interior relations of systems set up to give values to elements of structure and
the bits and pieces;
(3) the interior relations of contexts of situation.

1020 Context of Situation. The interior relations of the context of


situation may be summarized as follows:

A. The Relevant Features of participants: persons, personalities.


(i) The Verbal Action of the Participants,
(ii) The Non-verbal Action of the Participants

B. The Relevant Objects.


С The Effect of the Verbal Action. (112)

Individual situations are infinite in number and variety, typical situations


have not been defined structurally. Firth found Malinowski's Coral Gardens
and Their Magic (1937) was one attempt, Pareto's Sociology another, and gave
an offhand list of ten factors that might be profitably considered:

One method of tabulation would comprise ten entries as follows: (i) type of context
of situation; (ii) type of speech function; (iii) the language text and language mechanism;
(iv) the restricted language to which the text belongs; (v) the syntactical characteristics of
the text (colligation); (vi) other linguistic features of the text and mechanism, including
style and tempo; (vii) features of collocation; (viii) the creative effect or effective result;
(ix) extended collocations and (x) memorial allusions, providing serial links with preceding
or following situations. (Synopsis 1957:10)
10 Malinowski and Firth / 327

Firth is puzzled by Malinowski's attitude toward individual words, since


his ethnographic work presupposes and demonstrates that words are institut-
ionalized. Part of his critique might have been written by one of the
Modistae:

The descriptive linguist does not work in the universe of discourse concerned with
reality or what is real, and is not concerned with the ontological question of whether his
isolates can be said to have an existence or to exist — Let us again emphasize that facts
do not exist, they are stated, and it may indeed be a better guide to the handling of facts
to regard them as myths in which we believe and which we have to live with. (Synopsis
1957:113)

1021 Dictionary Definitions. Firth finds dictionary definitions are as


dangerous as they are useful because of their preoccupation with historical
values. Malinowski's analysis of freedom's multiple meanings in its universe
of semantic chaos' is 'perhaps the most interestingfull-length commentary on the
use of a common word Firth knew of, and occasioned remarks by Malinowski
that Firth thought to be of central importance (in Freedom and Civilization,
1947): all mental states which are postulated as occurrences within the private
consciousness of man are thus outside the realm of science (84), and We have
often stressed that in science we must run counter to linguistic usage. This is
even more important in social science than in the study of matter or organism.
(80) (1957.113-14) People do take fixed attitudes to words, but the physicist
does not inquire through universal suffrage or a Gallup Poll what the meanings
of his concepts are (81). These statements agree with Firth's idea, that
linguistic metalanguage refers chiefly to structures, systems and relations. Our
task is observation, analysis, synthesis, and renewal of connection. (114)

1022 Translation Meaning. Malinowski's contributions are most


important, but 'translation meanings, however systematic, do not in themselves
constitute linguistic analysis' (115). Recalling logic's de re and de dicto
distinction-wbat statements refer to, or statements themselves), Firth calls
attention to the distinction of use and mention of linguistic items. More
explicitly, he cites differences among (i) language under description, (ii)
language of description, and (iii) language of translation: (iii) subdivides into
(a) word-translation meanings and (b) translation meanings offered as a
means of identifying larger pieces, or as names for other native categories
supplied by informants (115). He finds Malinowski's warning worth repeating:

But there is nothing more dangerous than to imagine that language is a process
running parallel and exactly corresponding to mental process, and that the function of
language is to reflect or duplicate the mental reality of man in a secondary flow of verbal
equivalents. (Coral Gardens II.7; 1957.115)

1023 Linguists' translation Contributions. He invites comparison of


Malinowski's contributions with the results (UAL 1953) of a meeting of
linguists, anthropologists and philosophers. Levi-Strauss there outlined the
328 / GENERAL LINGUISTICS

problematic as: (1) relations between a language and a culture, (2) language
and culture, (3) linguistics as a scientific discipline and anthropology.
Discussions showed superficial agreement in terminology obscures rather than
fosters communication (116). He quotes from Jakobson's contribution on
norms of relevance changing within linguistics, mentioned above:

One of the most symptomatic features of this Conference was that we lengthily and
passionately discussed the questions of meaning... meaning remains a No Man's land. This
game of Give-away must end. For years and decades we have fought for the annexation
of speech-sounds to linguistics, and thereby established phonemics. Now we face a second
front: the task of incorporating linguistic meaning into the science of language. (1953.19,21;
1957.116)

Firth immediately appended to this, It is my personal opinion that


linguistics is suffering from a surfeit of phonemics and that our energies must
turn to the second front (117). In the same vein, he thought Hockett's
conclusion that ethnography without linguistics is blind; linguistics without
ethnography is sterile (1954.225) should read linguistics without meaning is
sterile', while agreeing with him that 'it had better be the linguists who work on
this systematic end of semantics (1954.250)~Firth's way:

...linguistics at all levels of analysis is concerned with meaningful human behavior in society
and that the structures and systems and other sets of abstractions set up enable congruent
statements of meaning to be made in exclusively linguistic terms... statements about
language data in terms of phonetics, phonology, grammar, stylistics, lexicography and
textual analysis in a background of statements of collocation and contexts of situation as
I understand those terms. (117-8)

In short, he finds Malinowski's contribution consists of sporadic comments,


immersed and perhaps lost in what is properly called his ethnographic
analysis. The two should be kept separate, but for later synthesis:

His outstanding contribution to linguistics was his approach in terms of his general
theory of speech functions in contexts of situation, to the problem of meaning in exotic
languages and even in our own. (118)

10.24 Structure and Structuralism in Firth. De Saussure anticipated a


paraphrase of Whitehead's 'Fallacy of Misplaced Concreteness': not what it is,
but what it contrasts with, constitutes a (linguistic) reality. Firth held
Phonemics neglected insights constituent of science (C) by overly stressing
columnar I-IV contrasts, thus neglecting their syntagmatic, contrastive reality.
Prosodic analysis was to study this neglected half of phonology.

1025 Implications. Firth chose to praise Malinowski as being no more


a structuralist than he. For him, that term referred to what he saw as obvious
omissions and confusions in Linguistics as practiced by Bloomfieldians, e.g.
relating phonemes as super-structures and allophones as alternations in
substructures'. His evidence was Malinowski's observation that
10 Malinowski and Firth /329
...the structure of all this linguistic material is inextricably mixed up with, and dependent
upon, the course of the activity in which the utterances are embedded. (Malinowski
1923.473; Firth 1957.101)

That one must study language structurally, Firth had no doubt, but he
found conceptual and practical confusions in the 'structuralist' grasp of
language 'structure'.

10.26 Actual and Potential Data. Saussure had defined la langue as pure
form, the intersection of two sets of relations he labelled associative (in
absentia) and syntagmatic (in prasentia). There is an observable empirical
relation between any item in parole, and other linguistic items with which it
co-occurs: these are actual, present, syntagmatic relations.
But even in parole, this concept of structure demands that there always
be potential, unrealized, empirically absent determinants which constitute the
valeur of what does occur. These are all items associated with it through
similarity or dissimilarity in form or content.
These are not empirical on a par with syntagmatic relations. Defining
them through positive as well as negative psychological associations makes
them empirically inaccessible.1 Basing them on formal or semantic similarities
and differences, makes them empirically remote. Hjelmslev proposed an axis
oi paradigmatic relations to replace subjective sets. This permits comparison
between attested forms and identifiable substitutes. (Actes du Quatrieme
Congres Internationale de Linguistes 1936.140).
Firth alluded to these distinctions in his popularization The Tongues of
Men in an example of syntagmatic implications among word-stresses in a
string like You know what I mean; You know what I mean: You know what
I mean; You know what I mean; You know what 7 mean; and You know what
I mean: ((1937), 1964)): each shift of stress has implications for the
pronunciation and interpretation of each member, as well as for interpretation
of the whole. These same facts are presented more technically by restricting
the terms system and systemic to a paradigmatic axis, while structure and
structural pertain exclusively to syntagmatic relations.

1027 Polysystemic Analysis. Language must be examined, he insisted,—


at all levels—on the basis of structures and systems. But there will be as
many systems to discuss as there are points of intersection that interest the
linguist. His polysystemic approach in phonology predictably yields different
elements and relations than the minimal pair test in phonemics, which Firth
found deficient because it was monosystemic.
Phonemics uses commutation (the minimal pair test) to inventory
contrasts that signal 'different in meaning'. The 'pairing" is systemic, abstracted
from structural setting. So contrast between CONvict and conVICT establishes
stress as a phoneme of English. But these forms contrast only in
metalinguistic use {Did you say CONvict or conVICT?). In We will conVICT
330 / GENERAL LINGUISTICS

the CONvict, they are in complementary distribution. Because phonemic


method prohibits 'mixing levels', syntactic phenomena are irrelevant at the
morphological or lexical level. On the principle, 'once a phoneme, always a
phoneme', this approach provides a single System of phonemic contrasts.
Pofysystemic analysis yields as many elements and systems as there are
relevant intersections. But relevant is defined in terms of the interests,
purposes, and criteria of the analyst. When more than the lexically distinctive
function of segments is relevant, different elements and combinations must be
'set up' to recognize and deal with them.

10.28 Collocation and colligation. Distinctions of system and structure


are not peculiar to phonology. Firth proposed the term collocation for
syntagmatic relations among lexical items, and colligation for syntagmatic
relations among grammatical categories. Both are forms of implication. The
constant involved in a formula for logical implication is a relationship,
expressible in English by If... then; variables which can fill that logical form in
the present discussion are lexical or grammatical items.
For collocations, Firth speaks of relations of 'mutual expectancy'
determining interpretation of pairs like sheer hypocrisy or mere detail. For
some speakers, hypocrisy and detail rarely occur without sheer and mere; for
others (and in language itself) their combination is an option. Colligations, on
the other hand, are intralinguistic demands among grammatical categories for
all speakers. For instance, in independent clauses, Attic Greek provides
options among indicative, subjunctive, and optative moods, corresponding to
actual, possible, or desirable viewpoints. In subordinate clauses, many
occurrences of indicative, subjunctive or optative are predictable as tied to,
demanded by, or colligated with, the tense of the independent verbs.
Implication2 is more general than relations in grammar or logic, and is not
limited to syntagmatic relations. Successions of events constitute syntagmatic
relations (Firth's structural axis). Their kinds are determined paradigmatically
(Firth's axis of system), such as past, present, от future; factual or contrary to
fact; actual or potential; successions by chance, or as consequences of laws
(natural, legal, moral, psychological, economic, political, social, logical,
associative, linguistic, etc.).
The notion of relevance here is again operative. Syntagmatic relations
hold among any actual succession of linguistic items by definition (as among
elements in ecology). But our control of the ecology progresses by describing
ecological systems, and explanation may result when higher systems are
identified. Refining the specific ways in which items in linguistic systems are
interdependent requires a linguistic abstraction. Firth labelled two such
abstractions colligation and collocation, because the domain of demand
relations is clearer among grammatical categories than the extent of toleration
among lexical items. Colligations are observable and predictable on the basis
of the language spoken; collocations correlate with what just happen to be
discussed in a particular language.
10 Malinowski and Firth / 331

10.29 Lexicon and Grammar. But if Firth's coinage of the collocation vs.
colligation distinction raises, without solving, the lexicon vs. grammar
distinction, it offers another resource for some perplexed grammatical usage.
For instance, particles called prepositions and prefixes were confused in ancient
grammars. Bloomfield distinguished them as free and bound forms. Later
grammars called particles governing (implicated with) cases of other free
forms prepositions, while bound particles called prefixes may not affect the
forms to which they are bound in the same way.
German 'separable prefixes' [Er ubersetzt (das Buch) ('He translates (the
book))'; Er setzt tiber (den Zaun) 'He jumps over (the fence)'], and English
'postposed prepositions' (The house I was telling you about; the mess we've
gotten into; much sought after, impossible to work with) seem to be halfway
between grammatical demands and lexical options. Either /or distinctions
distort three-term or more-term options.
In IE languages, the relation of the category Verb and the subcategory
Tense is colligation; in languages where time distinction is optional, it is
collocation if a word like yesterday suffices to express it. Where the category
Verb demands expression of knowledge vs. report (as in Quechua, Turkish,
and Acoma) the relation is colligation; in English, adding as I know or so
they say are phrasal, not lexical, and optional, not obligatory.

1030 Connotations. Collocations help distinguish connotations: in


American English, two professions predominate by 'habitual accompaniment'
(another of Firth's notes of collocation) by good—the clergy and doctors.
Journalists disarm suspicion of bias by referring to the good pastor or the
good doctor when their debatable acts or opinions are reported, never the
good lawyer or the good dentist. Occasionally one reads of the good Senator.
Collocations reinforce weakening bonds between sense and reference in
expressions like true facts, free elections, dedicated scholar, spirited charger,
real truth. Sports collocations with awesome, fantastic, dazzling superb,
professional, etc. rival theatrical commentary in relativizing nominal senses.

1031 Structural Conditioning. We have seen one kind of unilateral


conditioning of a linguistic form in Aristotle's addition of Quantification to
Plato's either/or definitions by dichotomy. A term like man can stand for
man-in-general, an indifferent number of men, or a single man. The number
of referents (extension) can vary while sense (intension, definition) remain
identical (rational animal). These are signalled in English by zero
modification (0/man) or addition of quantifiers like all, some or one.
We then saw that by Petrus Hispanus' time, mutual or bilateral
conditioning was recognized. He added qualitative modification to quantifiers'
ability to extend or restrict a term's extension in his discussion of mutual
influence among nouns, verbs, adjectives and restrictive relative clauses: while
man can be used to refer to anything meeting the sense rational animal, in tall
man, the word man refers only to tall rational animals; reciprocally, while the
332 / GENERAL LINGUISTICS

sense of tall by itself has to do with exceeding a norm, in tall man it refers
only to + norm-rational-animals.
Determination of colligations has always been the business, usually
dispatched rather readily, of grammarians: they involve closed sets. Lexical
sets are open, but woven together into the intelligible unity demanded by a
language's grammatical class meanings plus possible modifications. What
Firth contributed to this discussion was a way of incorporating some
extralinguistic factors in the context of speech situations that also help
determine interpretations.

1032 Prosodic analysis. Implications are to be investigated at all levels


of analysis. Firth saw axes of linguistic functioning at the intersection of
systems and structures: these are his technical terms for paradigmatic
(systemic) and syntagmatic (structural) relations:

S Y S S t r u c
T u r e EM

Studies in Linguistic Analysis (1957) contains Firth's own Synopsis of


Linguistic Theory and exemplification of its ideas in papers by eight other
scholars. Some terms clarify their perspective:

(i) The raw material for phonetics is referred to as the phonic material. From such
phonic material, phonic data are selected which may be described and recorded as
percepta by the techniques of phonetics. Selected phonic data, phonetically described,
are allotted to phonematic and prosodic categories, distributed in units, terms, classes,
structures, and systems.
(ii) A fundamental distinction is made between structures and their elements which are
syntagmatic, and systems of units or terms which commute and provide values for the
elements.
(iii) The theory of exponents, linking the phonic data phonetically described with
categories of phonology and grammar, is a central necessity of linguistic analysis at
congruent levels. (1957.vi-vii)

His phonology is not exhausted in phonemics' identification of systemic


distinctive units. He adds a construct explicating how sounds function as
structural elements as well. While morphophonemics involved grammatical
meanings by relating forms like singular wife, plural wives, or nominal mouth,
verbal mouth, that approach still assumes that its solutions must be stated in
terms of the single paradigmatic set of phonemes discovered by commutation
(the minimal pair test), or in terms of the distinctive features of which
phonemes consist.
10 Malinowski and Firth / 333

Some syntagmatic units (Firth's structural elements) have always been


recognized in labeb like syllable, consonant cluster, morpheme, word, phrase,
clause, sentence. If their ultimate constituents were morphemes, mediate
constituents could be morpheme-words, or words consisting of morphemes
with paradigmatic contrasts neutralized in syntactic analysis. No one would
quarrel with the identification of with all deliberate speed and quickly as an
adverbial function, despite the absence of an adverb in the former expression.
This phrase and adverb are in parallel distribution; the phrase is a substitute
for adverb, shares its class meaning and contributes its own constituent
meanings-adverbially (cf. Bloomfield 1933.251).
Syntagmatic units (syllables, consonant clusters) in phonemic phonology
are not established on the basis of distinctive lexical function. They are
frames within which the distribution of phonemes is conveniently stated. In
both, distinctive function is effected by phonemes, not their concatenation.
Syllables as units of rhythmic function are not comparable with the distinctive
function of phonemes. Plus juncture as phonemic in English was resisted by
those who failed to hear the fourth degree of stress it entails; others could
admit it as a potential clarification rather than an actual distinction on a par
with segmental contrasts. Clear instances are difficult to locate and probably
as rare as the potential confusions they are meant to distinguish.
What Firth intuited about phonology was that there are syntagmatic units,
or segmental units with syntagmatic implications of a phonetic nature, the
relevance of which has not been recognized because of the monosystemic
presuppositions of the phonemic approach. Linguists can agree that units
define each other negatively by contrast. So emphasis on their positive
composition (Firth's exponence) can be misleading. Firth distinguishes
between the phonic exponence of syntagmatic units he labeb prosodies and
their more abstract phonological contrasts and functions. A science is free to
define units within its own domain. It is neither falsified by, nor responsible
for, data defined to be tractable within another discipline. It is not surprising
that attempts to define prosodies in terms of 'what they are' {composition) as
opposed to what they do' (function) or 'where they are found' (distribution)
has caused confusion even among those convinced that he was on to
something important (cf. Matthews' review of Bazell et al. 1966). Their
function has to do with 'meaning in Firth's idiosyncratic grasp of that term.

1033 Phonemes and phonematic units. Firth labelled the phoneme a


transcribeme. This seems to relegate it to a preliminary, descriptive stage in
linguistics, as phonetics was demoted by phonological explanation. But Pike's
subtitle of his Phonemics as 'a technique for reducing languages to writing'
aptly describes the monumental contribution to a civilization that technique
makes. Firth had no intention of replacing phonemic transcription with his
non-transcriptional analysis. To call any stage of linguistics mere description
obscures the fact that any description is an implicit comparison and so an
implicit explanation.
334 / GENERAL LINGUISTICS

Like familiar phonemes, phonematic units are isolated by commutation


in minimal pairs. But phonic features (like the aspiration of thep's in popup)
are not abstracted from different places and then assigned to a single segment
taken to recur there.If popup's initial voiceless stop is aspirate and released,
its medial nonaspirate but released, and its final neither aspirate nor released,
there are three different systems of voiceless stops. Nothing a priori prevents
assignment of features like aspiration or tense articulation to more than one
phonological element such as both the phonematic unit p and the syllablepo.
Aspiration is a syntagmatic phenomenon, not just a paradigmatic or
segmental one. The breathiness of aspiration affects the first syllable of popup,
but is arrested by the second voiceless stop and does not persist throughout
the second and third syllables. While isolated [i] is usually assigned spread
articulation, the lip-rounding characteristic of English [r] persists throughout
the production of reek. It is a syntagmatic or structural—not just a
paradigmatic, systemic—factor. Such syntagmatic implications are prosodic.
These phonic facts can have functional implications. Firth recognized
phonetic meaning', since it is part of the 'meaning' of a native speaker to
sound like a native speaker. Pronouncing popup without its aspirate pattern,
or reek without rounded articulation, marks the speaker as a nonnative
speaker of English. For such reasons, Firth rejected the Phonemic label
redundant for aspiration or lip-rounding in these words as inappropriate.
Turkish words show vowel harmony. Adam ('man'), plural adamlar, and
dative plural adamlardan contrasts with ev ('house'), plural evler, and dative
plural evlerden. The fronted vs. back vocalic articulations of these words are
syntagmatic, structural contrasts, and their domain is the word, compared to
syllabic domains in popup or reeking. These contrasts are syntagmatic, not
paradigmatic. For Firth, the contrast is not redundant.

1034 IPA A useful contribution of Phonetics might have been to provide


a distinct symbol for every distinct sound in every distinct language of the
world. If this ever had been the aim of the International Phonetic
Association, it soon ceded to two practical determinations: The prohibitive
cost of type for such an alphabet, and the notion of the Phoneme as a
contrastive, rather than as a substantive unit.
The British distinction between broad and narrow (more or less phonemic
vs. refined phonetic) transcription is based upon two different facts about
language: phonetic transcription equips trained personnel to symbolize signals
identical in articulation to those of native speakers; accurate phonemic
notation equips them to produce the same number of distinctions as natives.
If phonological notation is based on sound articulatory analysis, its
symbols command articulations central to realizations of a language's
distinctive set of oppositions; if based on acoustic rather than articulatory
distinctions, the results are as problematic as what natives actually attend to.
In English, nondistinctive (hence, phonemically redundant) aspiration more
often distinguishes initial stops than the presence or absence of phonemically
10 Malinowski and Firth /335

distinctive voicing. If aspiration is omitted in I am very fond of cats, English


speakers tend to hear I am very fond of gats.
Empirically, voiceless stops, as segments, consist of silence. Phonemic
analysis of English might be said to attribute allophonic aspiration to silence
as to a segment (relevant to Firth's system) on a par with other continuants.
Spectrographic or kymographic analysis suggest that aspiration is a mode of
transition from silence to continuant (relevant to Firth's structure). No
phonemicist would deny that analysis.

1035 Prosodies vs. Phonemes. Phonemic analysis establishes contrasts.


Contrasts imply choice. Choice implies options, characteristic of phonemes.
Natives do not choose to produce or inhibit the automatic co-articulations
called allophonic in their choice of distinctive contrasts. Automatic, allophonic
articulations are part of the native's 'trained incapacities'. Natives attend only
to distinctive features when other natives speak. Most have to be retrained
to hear nondistinctive sounds in normal pronunciation, just as natives generally
notice only deviance from their own unformulated social conventions:
conformity is the normal and unremarked ambience in a community.
Professional retraining in how to listen can inculcate a phonetic or
phonemic attitude. Firth was a student and colleague of Daniel Jones, and he
of Henry Sweet, so attention to the positive, additive, data of phonetics was
more habitual to Firth than the negative, contrastive perspective of the
phonemicist. If the end-product of a contrastive notation is the goal, the
phonetic observation of the phonemicist can be preconditioned in terms of
paradigmatic contrasts. This is the basic goal of the minimal pair test.
When the end-product is to be holistic, i.e., both paradigmatic (Firth's
system) and syntagmatic (Firth's structure), phonetic observation is differently
conditioned. The phonological level is not even methodologically autonomous,
but viewed as one of a set of 'congruent' techniques to deal with holistic
Meaning. It attends e.g., to English aspiration syntagmatically and calls it a
prosody, it attends to the paradigmatic starting point of aspiration and calls
its exponence a phonematic unit. Use of technical terms like 'phonetic
substitution counters' instead of allophones or phonemes should not obscure
the role of the minimal pair test inevitably involved in discovering the requisite
segments, even in prosodic analysis, but does suggest that phonological
analysis need not to stop there.
Just what the function of structural (syntagmatic) units at the phonological
level is will depend on the language. It is plausible to assume that some
languages might not be illuminated by anything more than a phonemic
analysis (cf. Lyons 1962). Evidence from other languages suggests that
prosodic phenomena deserve separate statement.

1036 Non-lexical Distinctions. If, as the Terena analysis suggests to


Bendor-Samuel in Bazell 1966, grammatical functions can be realized prosod-
ically, then phonemic analysis alone can overlook facts. If, as Praguian
336 / GENERAL LINGUISTICS

analysis early suggested, functions of phonemes are demarcative and


culminative, not just distinctive, then a prosodic analysis of languages like
Turkish or Hungarian, where vowel-harmony marks out word separation, is
a neat addition.
The remote control explanation of retroflexion in Sanskrit (Allen 1951)
held that an earlier phoneme controlled the articulation of subsequent ones.
This was inexplicable empirically to some as assimilation or dissimulation,
because of the affected consonants were not contiguous, so was considered an
explanation repugnant to empirical sentiment. But to Firth it would be
aesthetically satisfying, and regarded as empirically acceptable, to postulate
retroflexion as a word-prosody with noncontiguous consonants within its scope.

Eugenie Henderson's 1948 Prosodies in Siamese: a study in synthesis


(Palmer 1970), often cited as a clear application of prosodic analysis, was
published the same year as Firth's foundational .Source and Prosodies (Palmer
1-26). Where Firth is theoretical and programmatic, Henderson is applied.
Her opening paragraph deserves attentive reading, since Firth monitored
papers of his Group, and this text can be taken to reflect precisely the thrust
of Firth's insights as then understood, but only as applied to Siamese (Thai):

The term prosodic feature is applied in this paper to certain properties


of modern spoken Siamese which may be regarded as abstractions apart
from the consonant and vowel systems. Such abstractions may be made
at the syllable, word, or sentence level. Syllable prosodies include tone,
quantity, and those prosodies which mark the beginning or end of a
syllable. Word prosodies include tonal and quantitative features, stress,
and the means whereby syllable is linked with syllable. Sentence
prosodies include sentence tone, and the means used to mark the
beginnings and end of phrases and sentences, and to connect phrase with
phrase or sentence with sentence. Italic type is used to show the
consonant and vowel units, and to name the prosodies, while heavy roman
type is used for phonetic transcription in general terms.

Notes

1. Inability to specify such associations publicly and objectively was pinpointed as a characteristic
of nonscientific mentality, in Bloomfield's review of Wundt's Vdlkerpsychohgie ((1913), Hockett
1956.39ff). Martin Gardiner (1983) identifies ability to link remote associations as the Art of Break-
through Thinking.

1. Implication derives from a metaphor of weaving strands into a totality; it suggests the image of
language as a net constituted by its intersections; it is lexically paraphrased in Collocations like:
"connected and interacting inferences about invo/vement of Accomplices implicated in the
accomplishment of a conspiracy deducible from the logical consequences deriveable from assumed
premises contained in the evidence". Matthews' 1968 review in Language of In Memory of J.R.
Firth deals with a study employing the Firthian notion of Collocation.
10 Malinowski and Firth /337

Reading

Firth, J.R. 1964. Tongues of Men and Speech. London. Oxford University Press. 221 pp. [Popular
style, but much quoted in Firth's Synopsis; outlines a program of Sociolinguistics.]
____ 1957. Papers in Linguistics 1934-1951. London. Oxford University Press, xii+233 pp.
____ 1957. Studies in Linguistic Analysis. Special Volume of the Philological Society. Oxford.
Blackwell. vii+205 pp. [Firth's Synopsis of Linguistic Theory 1930-1955; colleagues &
sympathizers exemplify or comment on his ideas.]
Malinowski, Bronislaw. 1923. Problem of Meaning in Primitive Languages, in Ogden & Richards
____ 1965. Coral Gardens and their Magic. 2 vols. Bloomington. Indiana University Press.
[Informative Introduction vii-xxii by Jack Berry.] Ogden, Charles Kay, and Ivor A.
Richards. 1969 (1923). The Meaning of Meaning. London.
Routledge and Kegan Paul. 10th edition, xxii+363 pp.

Supplementary Reading

Allen, W.S. 1951. Some prosodic aspects of retroflexion and aspiration in Sanskrit. Bulletin of the
School of Oriental and African Studies (BSOAS) 18.578-591. Bazell, C.E. et al., eds. 1966.
In Memory of J.R. Firth. London. Longmans, xi+500 pp. Bendor-Samuel, JohnT. 1966. Some
prosodic features in Terena. In Bazell et al. 30-9. Bursill-Hall, Geoffrey. 1960. Levels of
Analysis: J.R. Firth's Theories of Linguistic Analysis I &
П. Journal of the Canadian Linguistic Associations 6(2).124-135; (6(3).164-191).
_____1960. The Linguistic Theories of J.R. Firth. Thought from the Learned Societies of Canada
1960. Toronto. Gage. Firth, Raymond William, ed. 1957. Man and Culture: An Evaluation
of the Work of Bronislaw
Malinowski. London. Routledge and Kegan Paul. 292 pp. Bibliog. pp. 265-284] Gardiner,
Sir Alan. 1951. The Theory of Speech and Language. 2nd edition. Oxford. Clarendon. Lyons,
John. 1962. Phonemic and nonphonemic phonology: UAL 1962.28.127-33.
_____1966. Firth's theory of 'meaning', to Bazell et al. 288-302.
Matthews, P.H. 1968. Review of In Memory of J.R. Firth, ed. Bazell. Lg 44.2.(Part 1).306-17.
Mitchell, T.F. The Principles of Firthian Linguistics. London. Longman, xvii+213 pp.
Palmer, Frank R., ed. 1968. Selected Papers of J.R. Firth 1952-1959. Harlow. Longman.
x+209
_____1970. Prosodic Analysis. Harlow. Longman.
Robins, Robert Henry. Obituary of J.R. Firth. Language 37.191-9. In Sebeok II. 1967.543-54
_____1967. A Short History of Linguistics. Bloomington. Indiana University Press, vi+248 pp.
_____1970. Diversions of Bloomsbury: Selected Writings. Amsterdam. North-Holland.
Symmons-Symonolewiscz,K. 1958-9. Bronislaw Malinowski. Polish Review. New York. Arlington.
Temple R. 1899. A Theory of Universal Grammar as Applied to Savage Languages. The Indian
Antiquary vol XXVD. 197-208,225-35. Bombay.

También podría gustarte