Chomsky's Rebuttal of Whorf:
The Annotated Version by Moonhawk, 8/95
Noam Chomsky's Preface to Adam Schaff, Language and Cognition,
McGraw Hill, 1973. Translated from Polish.
of linguistic relativity as formulated particularly by Whorf2,
discussed here at length, is one that has given rise to much interesting
thought and speculation.
1 We have a problem already. You notice
there are no scholarly citations here, no page number to point out the
offending paragraph in Whorf's writings. Further, it assumes there is only
one hypothesis, which has been shown to be patently false because there
are as many hypotheses as graduate students and critics create. Someone
put it at 205. I dealt with five or six of the Hydra-heads (determinism
(weak & strong), nontranslatability, color terms, circularity, etc.)
in my "The Demise of the Whorf Hypothesis," in
which I argued for a moratorium on the offending phrase since only serves
to produce more heat than light, and has only tangentially to do with anything
Whorf ever wrote--which is why no one, even Chomsky, cites a page number
on such blatant mischaracterizations.
2 Whorf never in all of his writings
formulated a single hypothesis that I can find; he did however formulate
a principle, which is of a different order of scientific nomenclature than
a hypothesis--which Whorf knew, but has still 50 years later not become
clear to all "scientific" linguists. Again, no citation.
Many of the inadequacies
in Whorf's formulation are sketched here; there are others that deserve
more prominence than they have received.3
3 Well, the tone is certainly set, isn't
Whorf argues that
the structure of language4 plays a role in determining a world-view,5
and supports his argument by contrasting the6 world-view characteristic
of speakers of Standard Average European (SAE) with that of speakers of
various American Indian languages. As Schaff notes, the hypothesis practically
rests on the treatment of the categories of time and space7
4 Anyone who understands systems thinking
and then approaches Whorf has to conclude that Whorf saw language and culture
as a dynamic, interpenetrating system (e.g., p156). Chomsky is already
setting the stage here for arguing from a position of autonomous syntax,
and entirely different framing than Whorf used. Resetting the framing in
this way is a particularly successful rhetorical trick that Chomsky used
throughout his reign over linguistics, and he even told everyone about
it as he was describing what "others" (the media) do in order to "Manufacture
5 Notice very carefully the wording
here: the structure "plays a role in determining a worldview." Here Chomsky
is being very generous--is fairly accurate--and not overtly prejudicing
the argument to come. However, "determine" is a term unique to Newtonian/Cartesian
science, shorthand for "monocausally determines" such that event A uniquely
causes event B. Both the so-called "strong" version (which even critics
begrudgingly admit Whorf didn't hold, but anyone who speaks favorably of
Whorf probably does--although if Whorf didn't hold that view, why bring
it up?) and the so-called "weak" version (which most critics claim they
hold and figure Whorf maybe did too) nonetheless partake equally of Newtonian
monocausal determinism founded on an Aristotelian logic of exclusion. The
"weak" version is in fact more deterministic than any view Whorf held--which,
informed by the systems thinking of physics, partook of multicausal and
interdependent relationships founded on a non-Aristotelian logic of inclusion.
This is also behind the critics' distinction about whether language "shapes/molds"
thinking or merely "reflects/mirrors" our thinking. Systems thinking however,
with its insistence that sometimes the opposite of one profound truth is
another profound truth rather than something to be tracked down and killed--that
historically language shapes thinking and thinking shapes language. Notice
that Chomsky began by talking about linguistic relativity, and then all
of a sudden he's talking about determinism, as if the two have some inextricable
connection! Since, as I have shown elsewhere, Whorf's version of "linguistic
relativity" was an extension of Einstein's relativity principle--and Einstein's
relativity principle does not contain determinism. (Good thing his colleagues
didn't taunt IBM with "The Einstein Hypothesis!)
6 We might quibble on "the" or "a"
7 Without the careful use of quotation
marks, the way Whorf did, I'll be hornswaggled if those words space and
time there don't look like universal a priori notions given to all
humans in the same way!
The category of
space is similar in Hopi to SAE,8 but the Hopi, Whorf argues,
do not have our intuition of TIME as a smooth flowing continuum, with a
past, present, and future, in our sense.9 The basis for this
distinct world-view is provided by the categories of their language,10
which does not formally provide the past-present-future analysis of verb
forms, as in SAE.11
8 Except, of course for that pesky "imaginary
space" which our language and culture encourage us to have but Hopi language
and culture don't in Hopi speakers. This concept of "imaginary space" is
also extremely important for understanding these critical worldview differences.
(see esp. Whorf pp. 145-150).
9 Or, as Lakoff and Johnson elaborate,
the 3 popular metaphors of TIME AS A MOVING POINT, TIME AS A JOURNEY, or
TIME AS STATIONARY WHILE WE MOVE THROUGH IT. Now that we're talking about
metaphors instead of structure, we're into the culture part of the language/culture
system rather than the language (structure) part per se.
10 A skillful reiteration of
Chomsky's framing, whereas Whorf's statement (that "the Hopi language is
seen to contain no words, grammatical forms, constructions or expressions
that refer directly to what we call 'time,' or to past, present, or future,
or to enduring or lasting, or to motion as kinematic rather than dynamic....
or that even refer to space in such a way as to exclude that element of
extension or existence that we call 'time,' and so by implication leave
a residue that could be referred to as 'time.' (p57-8)") is preceded by
his own framing: "I find it gratuitous to assume that a Hopi who knows
*only the Hopi language and the cultural ideas of his own society* has
the same notions, often supposed to be intuitions, of time and space that
we have and that are generally assumed to be universal." Chomsky, coming
from a mathematical and autonomous syntax position, rather than an anthropological
and human-centered stance, must have somehow missed that critical 'language/culture'
linkage Whorf insisted on.
11 Actually, as I have shown
in a recent posting, "Summary: Time and Tenses" on the Linguist-List, neither
do most European languages! The future tense per se is unstable within
IE languages, and hardly occurs outside them, and the only reason we "feel"
that languages should have future tenses is because hundreds of years of
examining exotic languages through the lens of Latin Grammar told us so.
Against this it
has been argued that Whorf gives no evidence for a difference in linguistic
structure, but, rather, begs the question by postulating the difference
on the basis of the difference in the formal structure of Hopi and SAE.12
Here, then, is a point where further research might be proposed, perhaps
along lines that Schaff suggests, to bridge the gap in the argument.
12 Notice how effortlessly we move from
the relativity of concepts such as "space" and "time" to a different Hydra-pothesis
head called "circularity of argument."
But there is, after
all, a much more fundamental defect in Whorf's argument, namely, that his
description of SAE is incorrect.13 In English, for example,14
there is no structural basis for the past-present-future world view that
Whorf attributes, quite correctly, to SAE speakers.15
13 It's difficult to know here whether
Chomsky means that Whorf's description of "tense" or "time" in SAE is incorrect,
or his entire description of SAE (p134-159), including notions of naming,
plurality and numeration, nouns of physical quality, phases of cycles,
temporal forms of verbs, duration, intensity and tendency, habitual thought,
habitual behavior. Chomsky kind of leaves the question open.
14 But what kind of example is this--a
relevant one? In order to buy this example, do I have to also buy the assumption
that English is TYPICAL as far as SAE languages are concerned? What if
15 Here we go--slow down and watch
this one. Now Latin, which almost everyone would agree is more typically
SAE than English, does have a structural basis for that worldview, the
future tense, which Chomsky agrees is a correct description. English doesn't
have such a future tense--but rather than being considered untypically
SAE in this particular instance, it invalidates the entire SAE description.
Perhaps this rests on some claim by Whorf that I'm not aware of, that English
is the prototype of SAE languages.
Rather, a formal
analysis of English structure would show a past-present distinction, a
set of aspects (perfect and progressive), and a class of models, one of
which happens to be used to express future tense (among other devices that
serve this purpose). Approaching English from a Whorfian point of view16
we would conclude that an English speaker has no concept of time as a doubly
infinite line, he himself occupying the position of a point moving constantly
from past to future, but rather he conceives of time in terms of a basic
dichotomy between what is past and what is not yet past, in terms of an
aspectual system of a subtle sort, and in terms of a superimposed and independent
system of modalities involving possibility, permission, ability, necessity,
obligation, future (the latter not being distinguished in any special way).
The conclusion is absurd, which simply goes to show that our concept of
time is not determined by the linguistic categories,17 in any
detectable way, but is rather quite independent of them. If this is true
of speakers of English, why not of speakers of Hopi?
16 By which Chomsky means his own particular
radical refraining of Whorf's systems view to Chomsky's perspective of
autonomous syntax. Once you've distorted a person's position into a shredible
strawman, you can conclude anything you want to--a position Whorf strived
desperately to avoid (p59).
17 This is called "sleight of
hand" when a magician does it with physical objects. As noted in footnote
5, Chomsky set this up that Whorf "argued that the structure of language
a role in determining a worldview, [italic added], and "plays a role
in determining" cannot under any torturing of scientific nomenclature mean
"uniquely determines," which is the sense in "our concept of time is not
determined by the linguistic categories." If I read Chomsky correctly here,
he is claiming that linguistic categories play no role whatever in determining
our concept of 'time' and our worldview--again, a rather radical claim
which fits quite securely in an autonomous syntax perspective. Once we
allow Whorf his own framing, with language and culture as an interdependent
system, we may ask the question that Chomsky could NOT ask: given that
neither language system structurally supports a future tense, then which
language/culture system, English or Hopi, is immersed in centuries-old
metaphors of time as moving, stationary, river and journey?
of these matters one is led to several conclusions. First, the investigation
of linguistic relativity presupposes an exact analysis of linguistic structure
of a sort that is not available for SAE, let alone for American Indian
languages.18 This is no quibble over tenth-order effects. Even
an excellent linguist like Whorf was able to misconceive the nature of
such a basic part of English structure as the system of verbal auxiliaries.
What is more, it might yet turn out that Whorf's quite naive conclusion
about English is actually correct. That is, further research might, in
fact, show that at a deeper level of analysis than can be realized today,
there is a past-present-future system underlying the formal structure outlined
in the preceding paragraph.19 I see no indication that this
is true, but it would not be a very great surprise. If it turned out that
Whorf is correct, this would further substantiate my feeling that studies
of linguistic relativity are entirely premature, since his correct guess
would have been based on no evidence of substance and no defensible formal
analysis of English structure.20
18 This could just as well be said of
investigations into linguistic universals, of course.
19 Many linguists today are indeed
wondering who wrote in stone that a tense must be morphological, it is
on that basis ('will' is a modal rather than an affix) that Chomsky adjudges
20 When you get done with his preface,
what survives concerning the possibility that the Hopis do not have our
cultural notion of time? And that maybe, just maybe, there is something
profound we can learn about the diversity of human thinking? What happens
to that felt sense of difference that my American Indian friends have who
are bilingual? It just evaporates, becomes a non-issue, because of this
highly rhetorical style--which is the major problem I have with most or
all rationalist argumentation and why I, with Whorf, prefer systems thinking,
which doesn't just blanket-deny things but accepts all as a respected part
of the system, as the appropriate tool for.