What Whorf Really Said
by Nick Yee
Steven Pinker's main goal in The Language Instinct, as the title suggests, was to argue for the innateness of language acquisition. In a chapter titled, "Mentalese", he contends that words must be built upon word concepts, or what he calls mentalese, and not the other way around. His main opponent in this chapter was very naturally Benjamin Lee Whorf who Pinker claims is the primary proponent of the Sapir-Whorf hypothesis. Reading through his chapter, one cannot help but notice the vehement opposition that Pinker holds against Whorf's ideas. Pinker claims that Whorf's hypothesis and predictions are "wrong, all wrong" (Pinker, pg. 57) and that "the idea that thought is the same thing as language is ... a conventional absurdity" (Pinker, pg. 57). As one continues to read Pinker's chapter, one is inclined to align with his cause and conclude that Whorf was being too subjective and illogical. And yet, after some thought, one might begin to see the shallowness of Whorf's supposed hypothesis and wonder how such a backwards theory might have come to receive so much opposition if it's clearly wrong. One begins to wonder what Whorf really said.
Pinker defines the Whorfian hypothesis as having a weaker and stronger version. He defines the stronger version, known as linguistic determinism, as "stating that people's thoughts are determined by categories made available by their language" (Pinker, pg. 57). But Whorf never suggests such in his much quoted book of collected essays, Language, Thought, and Reality. In fact, Whorf defines "grammatical patterns as interpretations of experience" (Whorf, pg. 137). He sees language to be "in some sense a superficial embroidery upon deeper processes of consciousness, which are necessary before communication ..., and which also can, at a pinch, effect communication without language's and without symbolism's aid" (Whorf, pg. 239). Almost unambiguously, here is the mentalese that Pinker was talking about - thoughts that precede and are independent of words. So not only did Whorf anticipate mentalese, but he also clearly realized that language was not the same as thought which Pinker claimed as a conventional absurdity. Whorf goes on to say that language "may generalize down ... to something better - called 'sublinguistic' or 'superlinguistic' - and not altogether unlike ... what we now call 'mental'" (Whorf, pg. 239).
As for the weaker version of the Whorfian hypothesis, known as linguistics relativity, Pinker defines it as "stating that differences among languages cause differences in the thoughts of their speaker" (Pinker, pg. 57). Pinker then pushes this further and suggests that "the implication is heavy: the foundational categories of reality are not 'in' the world but are imposed by one's culture" (Pinker, pg. 57). And yet, Whorf realized that this was not the case and wrote that "there are connections but not correlations or diagnostic correspondences between cultural norms and linguistic patterns" (Whorf, pg. 159). Thus Whorf advocated a connection, not a correlation and certainly not a causality relation between language and cultural norms or cultural thought.
after setting up a flimsy strawman of Whorf's position, then proceeds
to describe how all of Whorf's evidence were either wrong or hoaxes. Pinker
first considers how Whorf claims that the Hopi language has "no words,
grammatical forms, constructions, or expressions that refer directly to
what we call 'time'" (Whorf, pg. 57). This is true, but what Whorf
meant was not time in general, but our concept of time. He actually spends
a whole chapter discussing the nonlinear properties of the Hopi perception
of time. Their aspectual and modal marker system allow them to refer to
events without using specific "time elements" and thus they
do not need to use tense markers. Whorf does not mean that the Hopi do
not understand what time is; he claims that the Hopi do not see time and
use tenses the way we do. Pinker interprets this as claiming the former
and happily pulls out an example where a native Hopi speaker uses a temporal
descriptor to counter Whorf:
Pinker also gives a short list (Pinker, pg. 60) of Whorf's Amerindian sentences and criticizes them on two counts. He first claims that the 3 sentence examples were from the Apache language, but that Whorf never studied Apache or had a true Apache informant. While it is true that Whorf never had an Apache informant or studied Apache intensively, the 3 sentences that Pinker produces are not Apache sentences, nor does Whorf label them so in his book. The first two sentences (boat and feast example) are from Nootka (Whorf, pg. 236 and 243 respectively), a language spoken in the Vancouver Islands, and the third sentence (gun example) is in Shawnee (Whorf, pg. 208). It is true however that Whorf studied these two languages. In short, Pinker is barking up the wrong tree.
Pinker also criticizes these sentences as being "clumsy, word-for-word translations, designed to make the literal meanings seem as odd as possible" (Pinker, pg. 61). This might be true, but Whorf did not intend for these sentences to be mere examples of Amerindian thought. He was trying to show specific aspects of the language in each example. For example, in the The boat is grounded on the beach example, the Nootka version is "It is on the beach pointwise as an event of canoe motion" (Whorf, pg. 236). Whorf meant for this sentence to demonstrate how Nootka does not use nouns, but rather verbs to express all events. Thus no "boat" is mentioned, only canoe motion that implies boat. In the He invites people to a feast, the Nootka version is "Somebody invites eaters of cooked food" (Whorf, pg. 243), where Whorf tries to demonstrate how the Nootka sentence is not divisible into the subject and predicate, or actor and object, relation that is essential in English.
At the end of this long ranting against Whorf, Pinker tells us "no one is really sure how Whorf came up with his outlandish claims, but his limited, badly analyzed sample of Hopi speech and his long-time leanings towards mysticism must have contributed" (Pinker, pg. 63). Now, although some of Whorf's examples were not entirely correct, such as the Eskimo words for snow, he was never entirely off the mark, and we have shown that many of Pinker's attacks are not valid. But even with the Eskimo words for snow, Whorf only suggested seven different words (pg. 210 and 216). Pinker speaks of "limited, badly analyzed samples", but one has to wonder whether Pinker or Whorf has made the grosser distortion of others.
After one has actually read Whorf's Language, Thought, and Reality, one notices that the better examples mentioned in his book are missing in today's "debate" over the Whorfian hypothesis. Most opponents pick on either how he was wrong with the number of Eskimo words for snow or how the Hopi do have a concept of time, or how mentalese must precede words. Yet Whorf had foreseen most of these problems and had included more detailed analyses as evidence. The most strikingly sustained one is in his comparison between Hopi and English.
Whorf begins his comparison at the very basic level of plurality and numeration. In English, plurals can be formed of both concrete and imaginary objects. For example, we can say "three dogs" when there are three dogs in front of us, but we can also say the next "three days" when the extra days are subjective and must be imagined in our minds. When we use "three dogs" however, the dogs we see may not be of the same size, but when we say the next "three days", the days must be of equal length in our imagination. In a sense, we count "three days" by recycling the one day three times in our imagination. Time, by itself, is a mere subjective "becoming later", and English speakers objectify it by dividing time into counted quantities, thus a "forward cycling". Now, one might argue here that this must be the case universally because that is the only way days can be perceived, and that the next "three days" is the same as the last "three days", one which is concrete and the other imaginary. Yet as Whorf states, "a likeness of cyclicity to aggregates [or objects] is not unmistakably given by experience prior to language, or it would be found in all languages, and it is not" (Whorf, pg. 139). In other words, if our way of perceiving and counting imaginary objects are so much a part of commonsense, then all languages should do the same. If they did not, then our way of thinking is not universally commonsense and implies that our language influences how we perceive imaginary objects such as time.
And this is exactly the case for the Hopi. They can only count concrete nouns, such as dogs, and cannot do so for imaginary nouns. For imaginary nouns, they have to combine the ordinal with singulars. So to say, "They stayed ten days", the Hopi have to say instead "They left on the eleventh day". What this does is that it leaves the subjective quality of time intact - things happening "later and later" instead of things happening "in cycles" or after "a certain number of cycles". Thus, "instead of our linguistically promoted objectification of that datum of consciousness we call 'time', the Hopi language has not laid down any pattern that would cloak the subjective 'becoming later' that is the essence of time" (Whorf, pg. 140).
Whorf then uses this background as a stepping stone to more complex linguistic rules. He next considers count nouns and mass nouns. English makes a distinction between count nouns and mass nouns. Count nouns are those objects that are discrete and have clear boundaries - dogs, cats, pins, bottles. Mass nouns are those that are continuous or have no clear boundaries - salt, water, clay, dust. To count mass nouns, we need a container word. "Hence the 'lumps, chunks, blocks, pieces,' etc., seem to contain something, a 'stuff', 'substance' or 'matter' that answers to the 'water', 'coffee', or 'flour' in the container formulas" (Whorf, pg. 141). A lot of English is built upon this complementarity of formless item and form. Hopi, however, does not have mass nouns. All nouns are countable. Nouns for water are still indefinite, but do not imply lack of outline or size. Thus, they do not have a need to think of objects as the interplay of formless items and forms.
The implications of this are more clearly seen in our model of time. Time for us is a mass noun and thus it has to be quantified by container nouns, such as "an hour of time", "a moment of time". We cannot say "one time", or "two times" the way we want it to mean. For us, time is carved into quantities by our "summers, winters, Septembers, and noons." For the Hopi however, phase terms, such as "morning" are almost like adverbs, not nouns. "Summer" is only when it is hot and dry. Our "It's summer" translates more readily into Hopi as "It's hot and dry right now". Thus there is no "This summer" in Hopi, only "Summer now", "Summer recently". Because they have no form containing a formless item, they have no need to quantify "time" into cycling days or summers.
So for English, it is convenient to separate time into past, present and future because of these subdivisions of time. And to distinguish between these three time frames, we use tense markers. For the Hopi, however, they need only have two - the experienced and the yet to be experienced. What is important to the Hopi is the validity of a sentence and they use validation markers instead of tense markers. A comparison between the two systems is given:
Thus, the Hopi are able to describe all phenomena in the world without ever having to recourse to a notion of linear time like English speakers do. Whorf then mentions the effects that such different systems have on the two cultures. The Hopi see tomorrow as a reoccurrence or revisit of today, rather than a new part of a different cycle. This is why they believe in the accumulation of rituals or preparations because they carry into the reoccurrence and do not disappear. Another striking difference is that because they have no need to count imaginary nouns, they believe that when one thinks of a rosebush, one's thoughts are actually interacting with the real rosebush, not a mere mental one. For English speakers, the impact is seen in the premium we place on time: "time wages, rent, credit, interest, depreciation charges, and insurance premiums" (Whorf, pg. 153). It also explains the monotony and regularity impressed on English-speaking cultures.
But if Whorf never advocated linguistic determinism or linguistic relativism, what was this sustained example for? What was it doing in his book? Certainly, this example seems like a salient point to support either the weak or strong version of the Whorfian hypothesis.
In fact, Whorf himself answers these questions many times over in his collection of manuscripts. What Whorf claimed was not a linear system where either language influenced culture, or culture influenced language, but where "in main they have grown up together, constantly influencing each other" (Whorf, pg. 156). As in the sustained example between English and Hopi, one is hard-pressed to say where the linguistics begin and where the culture ends. If commonsense about numbers and counting exist, why didn't both English and Hopi develop similar counting schemes. And when did the language influence the culture back? Whorf admits that we can't really tell, "but in this partnership the nature of language is the factor that limits free plasticity and rigidifies channels of development in the more autocratic way ... [language] can change to something really new only very slowly, while many other cultural innovations are made with comparative quickness" (Whorf, pg. 156).
The individual to Whorf is a microcosm that contains "all the give-and-take between language and the culture as a whole, wherein is a vast amount that is not linguistic but yet shows the shaping influence of language" (Whorf, 147). In this dynamic system are the variables of thought, language and culture which feed off each other and create the "microcosm that each man carries about within himself, by which he measures and understand what he can of the microcosm" (Whorf, pg. 147).
For Whorf, the whole picture is necessarily complicated. Our thoughts and perception must have impact on our words. Whorf suggested that:
For instance, if a race of people had a physiological defect of being able to see only the color blue, they would hardly be able to formulate the rule that they saw only blue. The term, blue would convey no meaning to them, their language would lack color terms, and their words denoting their various sensations of blue would answer to, and translate, our words 'light, dark, white, black' and so on, not out word 'blue'.
But words form part of grammar, and "formulation of ideas is not an independent process, strictly rational in the old sense, but is part of a particular grammar" (Whorf, pg. 212). Culture is built from an accumulation of thoughts and words, which in turn has a give-and-take relationship with language as Whorf articulated in quotes above. And in turn, language affects how we "cut up and organize the spread and flow of events as we do, largely because, through our mother tongue, we are parties to an agreement to do so, not because nature itself is segmented in exactly that way for all to see" (Whorf, pg. 240). And we have come full circle, but in the real world this cycle never ends.
Whorf answered the current debate on the Whorfian hypothesis almost 50 years ago in his book. As to the strong version, Whorf would claim that it is not the case that language determines thought categories, because this phrasing reduces the dynamic transformation into a simple one-way linear state. As for the weak version, Whorf knew that while there were connections between language and culture, there were no correlations. One could not say that languages that have no tense markers do not objectify time and thus have cultures that believe in ritual accumulation. And a case in point would be Chinese, which has no tense markers, but whose nouns are all mass nouns, and which certainly objectifies time. It is fascinating that Chinese can get along without tense markers even though we have the Western view of time being cyclical, and can indicate past, present and future without recourse to tenses. This further demonstrates how intertwined culture, thought, and language are. It cannot be that noun classes determine tense classes, or Chinese, like English, should have tense markers while Hopi have none.
Furthermore, Whorf probably never intended for gross language differences such as: 1) lexical differences (one culture has a word for "filial piety", while another doesn't), or 2) syntactic differences (one culture has a counterfactual construction while another doesn't) to connect with cultural or thought differences. The examples he gave were more about how a commonsense is hidden in basic grammar, such as that of the Hopi and English comparison.
An example from Takao Suzuki's book, Words in Context, is truly in the spirit of this kind of Whorfian example. And it is an English one. Consider the verb "to drink" in English. You can drink coffee, tea, water, and soup. So it appears that drink has to do with liquids and it has to do with not chewing whatever is ingested. But we do not drink all liquids. For example, lighter fluid is "fatal if swallowed", not "fatal if drunk". This is true for all poisons as well. Also, we "swallow" solid medicine, but we "take" liquid medicine. So we see that in two special cases - poisons and liquid medicine, we do not use the word drink. One may now assume that drink means "orally taking some liquid that is expected to maintain one's physical well-being" (Suzuki, pg. 19). But even this misses something, because we can say "He drank vodka till he passed out", and here we know that drinking tequila does not exactly maintain your well-being. Another clue to what drinking really refers to is seen in Socrates' death. How did Socrates die? People usually say that "He died by drinking hemlock", not "He died by swallowing hemlock", and here I think we see more clearly that drinking must also have a volitional aspect to it - the wanting to ingest something. So drinking is really "the voluntary oral ingestion, without mastication, of liquids, excluding liquid medicine, with which that one expects to, whether correctly or not, reach a specific physical state in mind." I really don't think any dictionary or native speaker would define drink in that way. And that is exactly Whorf's point.
Because we cannot accurately define such a common word, we are not even aware of the logical boundaries it imposes on our formulation of ideas. When we say drink, we are actually invoking personal volition and end-state goal. When we need to use the word drink, our mind must be logically filtering in and out the necessary conditions: Is chewing involved? Is it liquid medicine? Is it voluntary? Is there a goal of physical state in mind? All this logical equating, however, is invisible to us. All we say is the word "drink", and when it's used inappropriately, we feel this awkward nudge inside of us, but we can't articulate why it's wrong. And "drink" is just one word. Imagine the net of logical if/or/and's that language must impose on us that we are not conscious of. This is what Whorf truly meant when he says that "formulation of ideas is not an independent process, strictly rational in the old sense, but is part of a particular grammar" (Whorf, pg. 212). Language contains commonsense.
And yet, we should be painfully aware of the interplay between thought and language - that someone must have created the word from a concept, but that the word has been shaped by culture and time, and it is now unconsciously shaping our thoughts.
This paper began with Pinker's stand against Whorf's ideas. Not only did Pinker distort many of Whorf's ideas however, we have seen that contemporary discussions of Whorf never articulate the dynamic system he suggested. In fact, most of the debate that comes under the heading of the Whorfian hypothesis has nothing to do with Whorf himself. It might surprise many that Whorf never uses that term in his book. In the places where he refers to it directly, he calls it the "linguistic relativity principle" (Whorf, pg. 214 and pg. 221). A principle is like a theory that tries to explain phenomena in the environment, whereas a hypothesis is a prediction based on an underlying theory. Whorf saw his theory as more of an intrinsic part of reality from which hypotheses could be made, rather than a hypothesis itself. He was trying to explain phenomena rather than trying to predict one.
Contemporary accounts of the "Whorfian hypothesis" also neglect another important element that recurs in his papers, and that is his motivation for carving out such a principle. The impetus lay mainly with the infant science of quantum relativity and quantum mechanics in the 1940's. Scientists realized that as they entered the quantum realm of particles and quarks that normal words could no longer correctly describe what was happening, and that illogical things happened constantly. Even half a century later quantum physics still holds very counterintuitive concepts and notions. R.P. Feynman has claimed that "I think I can safely say that nobody today understands quantum physics" and Roger Penrose (1986) commented that the theory "makes absolutely no sense". Whorf realized that this was due to the way our language was structured. We had taken common sense too far and are surprised that "Newtonian space, time and matter are no intuitions. They are recepts from culture and language. That is where Newton got them" (Whorf, pg. 153). He notes that "modern thinkers have long since pointed out that the so-called mechanistic way of thinking has come to an impasse before the great frontier problems of science" (Whorf, pg. 238).
We are confused for example by experiments that show interference of electrons where only one was used. In layman's terms, that is like a one-handed man clapping. Whorf suggests that we have problems with this counterintuitive situation because our grammar forces us to construct sentences and in turn ideas into actor-action pairs where events are caused by something else. Even in cases where there is no real actor, actorship is assigned to a dummy pronoun, such as "It is raining", where "it" is referring to nothing but which makes it sound like the "it" is doing the raining. But in Nootka, as Whorf showed in the boat and feast example that Pinker dismissed as clumsy, they do not have subjects or objects, but rather just verbs that imply what we call nouns. Action occurs independent of subject or object, and it is not surprising that many Amerindian speakers have much less trouble comprehending the paradoxes of quantum mechanics than English speaking scientists do.
What Whorf wanted to do was to suggest that a better grasp of the modern physics could be achieved if we made a conscious effort to make our physics language more verby, and less agent-oriented, and this is an approach that David Bohm has suggested. Whorf also uses this as an occasion to explain why other less technological cultures and languages may help technological advancement in ways that could not have been foreseen.
It feels appropriate to end this discussion of what Whorf really said with his most often-quoted passage that most undergraduate Psychology, Linguistics and Anthropology students read somewhere. Many of us have read the passage before, but it is hard to read that passage in the same light again:
We dissect nature along lines laid down by our native languages. The categories and types that we isolate from the world of phenomena we do not fine there because they stare every observer in the face; on the contrary, the world is presented in a kaleidoscopic flux on impressions which has to be organized by our minds - and this means largely by the linguistic systems of our minds. We cut up nature, organize it into concepts, and ascribe significances as we do, largely because we are parties to an agreement to organize it in this way - an agreement that holds throughout our speech community and is codified in the patterns of our language. The agreement is , of course, an implicit and unstated one, BUT ITS TERMS ARE ABSOLUTELY OBLIGATORY; we cannot talk at all except by subscribing to the organization and classification of data which the agreement decrees. [emphasis from original] (Whorf, pg. 213-214)
His first few lines seem so deterministic that one does not pay much attention when he talks about the origins of this agreement. The agreement "holds throughout our speech community and is codified in the patterns of our language". Because it is codified in our language, he cannot be referring to language itself, but must be referring to culture. Thus, it is cultural and social mores that are codified in our language, which in turn dissects nature. Wherever this quotation appears, we are often so primed by the opposition to the "Whorfian hypothesis" that we read Whorf as being completely deterministic when he is actually subscribing to a dynamic, fluid and ultimately holistic viewpoint.