Might the language we speak affect the way we think? The very idea that language could determine the nature of our thinking carries more than a faint whiff of anachronism. The notion seems to belong to an altogether different age, prior to the serious study of mind as an information processing unit. In this era where an MRI lights up regions known as Broca’s and Wernicke’s area, language seems to be completely understood right down to the neurons of the arcuate fasciculus, the fiber tract connecting these two "language regions."

Yet although this thesis of linguistic determinism seems at first sight to be outdated, the idea can easily shed light on modern theorizing in ways that make contemporary language models look anything but silly. The possibility that the language we speak influences the way we think has excited both popular and scientific imagination in the West for well over a century (Lucy, Rethinking, 37). Many theories have attempted to reconcile language and thought by means of environmental and cultural influences. Amid this myriad of speculation, hypotheses, and cultural doctrines, however, none is more popular and subject to scrutiny than the Whorfian theory of linguistic relativity. 

WHORFIAN THEORY DEFINED

The current Whorfian hypothesis has strong roots in German educator Wilhelm von Humboldt’s study of linguistic relativity and determinism early in the last century (Slobin 70). According to Humboldt languages differ from one another; thought and language are inseparable; and, therefore, each speech community embodies a distinct world-view. 

Benjamin Whorf extended this doctrine of linguistic determinism to describe the roles of language and thought in human development. Bringing the idea to a new and heavy mix of empiricist epistemology, Whorf placed an emphasis on the unconscious influence of language on habitual thought:

Actually, thinking is most mysterious, and by far the greatest light upon it that we have is thrown by the study of language. This study shows that the forms of a person’s thoughts are controlled by inexorable laws of pattern of which he is unconscious. These patterns are the unperceived intricate systematizations of his own language—shown readily enough by a candid comparison and contrast with other languages, especially those of a different linguistic family. His thinking itself is in a language—in English, in Sanskrit, in Chinese. And every language is a vast pattern-system, different from others, in which are culturally ordained the forms and categories by which the personality not only communicates, but also analyzes nature, notices or neglects types of relationship and phenomena, channels his reasoning, and builds the house of his consciousness.  (qtd. in Gumperz and Levinson 21).

Linguistic relativity is a theory primarily about the nature of meaning, "the classic view focusing on the lexical and grammatical coding of language-specific distinctions," as Gumperz and Levinson describe (7). In Whorf’s theory, two languages may code the same incidents utilizing semantic concepts or distinctions peculiar to each language. As a result, each language reflects different perspectives of the same bit of reality.

The Whorfian hypothesis can therefore be summarized as follows (Gumperz and Levinson 25):

(1) Different languages utilize different semantic representation systems which are informationally non-equivalent (at least in the sense that they employ different lexical concepts); 

(2) semantic representations determine aspects of conceptual representations;

therefore

(3) users of different languages utilize different conceptual representations.

Many authors find this thesis of linguistic determinism wildly adventurous or even ridiculous. Linguist and published author Steven Pinker has stated, "The discussions that assume that language determines thought carry on only by a collective suspension of disbelief" (qtd. in Gumperz and Levinson 23). Cultural linguists Wason and Johnson-Laird criticized Whorf’s hypothesis of "engendering much confusion and many circular arguments" (23). According to this team of researchers, "there is no evidence for the strong version of the hypothesis—that language imposes upon its speakers a particular way of thinking about the world" (23). 

Nevertheless, many other linguists have recently claimed to find the Whorfian hypothesis plausible and extremely advantageous in studying the role of language. New studies now offer the strong evidence that Wason and Johnson-Laird had requested. Researchers studying the linguistic relativity hypothesis have purported that the differences among languages in the grammatical structuring of meaning influence habitual thought (Lucy, Grammatical Categories, 1). A rigorous demonstration of such influences would have profound implications not only for the scientific understanding of human life but also for the conduct of research and public policy. The results of the Whorfian hypothesis can perhaps illuminate our very thoughts in ways no MRI nor neurobiology text has yet to explore.

LINGUISTIC RESEARCH

Taking Whorf's formulation oand subsequent empirical research into consideration, four components of adequate empirical research must be present for an accurate, effective analysis of different languages (1):

1. First the research must be comparative; that is, it must present contrastive data on two or more language communities. Without a comparative dimension, nothing can be established about the impact of language differences. Ideally, the languages at issue should contrast widely so that clear, strong differences are at stake.

2. Such comparison should take an external non-linguistic reality as the metric or standard for calibrating the content of linguistic and cognitive categories. The development of an adequate, neutral metric represents the crucial analytic problem in any project on the relativity issue. Without such a metric, one cannot say with precision in what respect one language’s categories differ from another’s or establish that a language category and a cognitive category concern a common domain.

3. The languages of the communities being studied must be contrasted as to how they differently construe a common reality. Such an analysis should focus on a whole configuration of meaning rather than on an isolated lexical or grammatical category, and it should attempt to state the contrast between the configurations within a unified framework.

4. The implications of the language differences for thought must be articulated. This involves proposing plausible cognitive entailments of the habitual use of the language patterns at issue.

CROSS-CULTURAL WORDPLAY

Cultural linguist Franz Boas catalogued a great diversity of obligatory grammatical categories across languages in the introduction to his Handbook of American Indian Languages (Slobin 71). For example, Boas discussed the English sentence, The man is sick, and noted that in Siouan one would have to indicate, grammatically, whether the man is moving or at rest (71). In Kwakiutl one would have to indicate whether the man in question is visible or non-visible to the speaker, and near to the speaker, hearer, or a third person. In Eskimo, one would simply say, "man sick," with no obligatory indication of definiteness, tense, visibility, or location. Boas’s concept can be extended to other languages such as Spanish where one must indicate whether the man is temporarily or chronically sick. In many other European languages one cannot indicate definiteness apart from gender (71). As a result, Boas stated:

It will be recognized that in each language only a part of the complete concept that we have in mind is expressed, and that each language has a peculiar tendency to select this or that aspect of the mental image which is conveyed by the expression of the thought (qtd. in Slobin 71).

DANGEROUS THOUGHTS

Yet does a linguistic focus on a particular image actually change our perception of the "larger picture"? Linguist Paul Kay suggests that "it is a dangerous jump from the observation that two languages provide different ways of talking about a given subject matter to the conclusion that the speakers of those languages think of that subject matter in distinct ways" (97). In Brazil the limited-access highways are marked with the word entrada plus the name of the locality, while those of the US may be marked with the word exit plus the name of the locality. Thus, we can compare Entrada São Jose versus San Jose Exit. Kay stated, "Such an observation . . . might lead a Whorfian dealing with two very different cultures to spin a tale about how members of the cultures speaking these languages had distinct conceptions of the nature of road systems, while it is perfectly obvious that Brazilians and North Americans have exactly the same understanding of freeways" (97). In English one speaks of screwdriver, in German one speaks of schraubenzieher (screw puller), and in French of a tournevise (screw turner). Wouldn’t it be folly to infer differences in thought or cultural conceptualization from such differences in naming? (98).

Kay proposes that the Whorfian hypothesis can only be useful for analysis within a culture opposed to cross-cultural comparison. Consider, for example, the words strictly speaking and technically, both commonly used within the English language. If it is accepted in the following examples that Sacco and Vanzetti were unjustly convicted, (1) is true and (2) is false (100). Thus different truth judgments for the same situation arise depending on these variations in semantic framing. 

(1) Technically Sacco and Vanzetti were murderers.

(2) Strictly speaking Sacco and Vanzetti were murderers.

Kay concludes that "linguistic relativity seems to be as much an intra-individual matter as an inter-cultural one" (101). If linguistic relativity so reconstructed is, however, a language-internal phenomenon as much as a cross-linguistic one, then its consequences for inter-cultural communication, and so on, may be less dire than often supposed. In pursuing a deeper understanding of linguistic relativity, Kay reminds us to be aware "of the large amount of alternative schematizing of a given event . . .within each language and hence within each speaker of that language" (111). 

Cultural linguists Bivens and Berk observed and recorded the ongoing speech of a group of children between the ages of 5 and 10 (qtd. in Clark 195). They discovered that the incidence of private speech increased when the child was alone and trying to perform some difficult task. In subsequent studies, the researchers learned that those children who made the greatest numbers of self-directed comments were the ones who mastered the tasks best. Hence, Bivens and Berk concluded from this study that self-directed speech is a crucial cognitive tool that allows us to highlight the most puzzling features of our environments (195). The Whorfian hypothesis can therefore be argued to exist on a personal level.

A PINWHEEL OF COLORFUL NAMES

Many psycholinguists have focused on lexical items, especially ones for colors, to demonstrate that Whorf’s theory can be applied to a cross-cultural study. In one of the most famous of these studies, Brown and Lenneberg tried to show that certain colors were more "codable" than others in English (qtd. in Lucy, Rethinking, 45). Subjects assigned them shorter names and tended to agree more on the application of those names to color samples. The more codable colors were recognized and remembered more readily than the other colors (45). 

Extensions of the early color work by anthropologist Berlin and his collaborators generated the first broad multilanguage comparative framework to actually be applied to the relativity question (46). Zuni, a language of the American Southwest, for example, exhibits two terms that we might translate as "yellow" (46). Closer analysis reveals that one term is verbal and refers to things that become yellow by ripening or aging whereas the other is adjectival and refers to things that have had yellow substances applied to them. The customary approach in Zuni would select one term as "basic" and ignore the aspect of its meaning (i.e., the manner of becoming colored) for which there is no English equivalent (46). Hanunóo, a language of the Philippines, has four terms that seem to refer to what we would call white, black, green, and red but which under further analysis turn out to mean roughly "lightness, darkness, wetness, and dryness" (46). 

Although this work has been highly criticized for its assessment of relativity, the study has shown that cultures interpret colors differently as a result of their languages. While some cultures may associate color names with tactile touch, others link the color names to internal development (aging, ripening, etc.) Hence, the cross-cultural pinwheel of color linguistics has demonstrated that the grammatical structure of language can influence our thoughts and interpretations. Brown, Lenneberg, and others involved in this study also inaugurated a tradition of assessing thought by presenting individual subjects with experimentally controlled memory tasks rather than analyzing naturally occurring patterns of everyday behavior (47). Such procedures played an equally important role in Dan Slobin’s storybook experiment.

A PICTURE’S WORTH A THOUSAND WORDS

Consider the pictures in Figure 1. 

Both are pages from a picture storybook that were used by Slobin and his colleagues to assess the Whorfian hypothesis on a cross-cultural perspective.  These pictures represent a pair of events that you can understand immediately, probably without talking to yourself at all (Slobin 72). Something happens to the boy in the tree, and something happens to his dog. An owl and some bees are involved; the location is most likely in a forested area. What grammatical categories are present?

As an English-speaker, it will be evident to you that the activity of the dog is durative, or extended in time, in comparison to the activity of the boy (72-73). In narrative mode, you might say, "The boy fell from the tree, and the dog was running away from the bees." English marks a progressive aspect on the verb, seeming to correspond to an obvious temporal component of the "complete concept" or "mental image" (73). If you are a Spanish-speaker, you, too, will recognize the durativity of running, because Spanish also has a progressive aspect, as well as imperfect aspect. Nonetheless, this speaker might also note that the falling of the boy is punctual or completed, since Spanish makes a contrast between perfective and imperfective aspect. However, what if you speak a language that has no grammatical marking of perfective/imperfective or of progressive, such as German or Hebrew? (73). Linguist Boas would presumably have suggested that you are aware of the differences in temporal contour between falling and running, but simply have no need to mark them grammatically in your language. 

Consider the following descriptions of Figures 1 given by two five-year-olds. One is in English and the other is an English translation of a Spanish interpretation, using the progressive to correspond to the original imperfective. In subsequent pages, the boy accidentally gets entangled in the antlers of a deer and falls into some water with his dog (responses excerpted from Slobin 77).

First response: The boy looked in a hole in the tree. An owl came out that threw the boy. And the dog, the wasps were chasing him. The boy hid behind a rock and the owl flew away. A deer that was behind the boy when he climbed . . . And he slipped on top of the deer, while the deer was running. The dog went first. He threw them down where there was a river. Then he fell.

Second response: And the boy looked in the tree. And then the boy fell out, and the owl was flying, and the dog was being chased by the bees. And then the boy got up on some rocks, and the owl flew away. And the boy was calling for his frog on the rocks. And a deer . . . the boy got caught on the deer’s antlers. And then the deer carried him over a cliff and threw him over the cliff into a pond. And the boy and the dog fell, and they splashed in some water.

The two versions are similar in their treatments of movement through time (78). Both narratives mark some events as being in progress. In the first, compare: threw vs. was running; in the second, fell vs. was flying. English and Spanish both have these markings of durativity.

The two versions differ, however, in their treatment of location and movement through space. In the first version, trajectories are not highly elaborated: threw the boy, slipped on top of the deer, threw down. The second version depicts more detailed trajectories: fell out, carried over a cliff, threw over the cliff into a pond, splashed in some water. On the other hand, the first version has relative clauses that depict static locative configurations, which are lacking in the second (78). Notice, for example, the child from the first responses says: a deer that was behind the boy, where there was a river. 

These cues, alone, are sufficient enough to differentiate the Spanish speaker from the English one. Where English allows for elaborated trajectories of motion, Spanish has simple verbs of change in location, supplemented by more elaborated descriptions of static locations of objects. By providing a perfective in addition to imperfective and progressive, Spanish makes it possible to grammatically mark both poles of the durative-nondurative distinction, whereas the English progressive provides explicit marking only of the durative pole (79). 

German and Hebrew lack distinctive marking of either pole of aspectual contrast (79). Hebrew has no grammaticized aspect at all; verbs are simply inflected for past, present, or future tense. German has a simple past and present. Neither language has grammatical marking of either progressive or imperfective. Consider table 1 summarizing these differences.
 

	Preschool	School	Adult	OVERALL
Hebrew	71	100	63	78
German	54	80	78	81
English	26	22	33	27
Spanish	23	18	0	21

 

Table 1: Percentage of narrators using same tense/aspect form for "fall" and "run" clauses from Figure 1 (Slobin 80).

PLURALIZING THE DIFFERENCE

John Lucy also applied the Whorfian hypothesis to his research on the grammatical treatment of nominal number (for example, pluralization) in Yucatec Maya and American English. Lucy further argues that these differences correspond with detectable differences in habitual thought as assessed through simple cognitive tasks involving attention, memory, and classification (Lucy, Grammatical Categories, 3).

Background on the Yucatec Maya Culture

The Mayan people studied by Lucy reside in a small rural village in the state of Yucatan in southeastern Mexico. Historically, contact with the outside world for the villagers required a day’s journey by foot or on horseback (8). During such a journey, one would encounter scrub jungle, small individual ranches, and farms belonging to private landowners. The village is called an ejido, a place where residents and their descendants have petitioned for and been given land collectively by the national government. In rotation, men serve as head (comisario) of the ejido organization and the men who farm in the village serve as a committee of the whole in making decisions about matters involving the community. Village life centers on corn farming and associated religious practices (12). 

Maya is the language of the village; only a handful of men are truly bilingual. Education in the village is of two varieties: a formal education in a school system which sends in teachers each week and an informal one provided by parents and by general exposure to life in the village (15). Lucy chose to study such a society since "cultural conservatism was deemed important to minimize complicating factors associated with modernization" (17).

Experiment

Lucy chose to focus on nominal number markings as a means for testing the Whorfian hypothesis. Notionally, this term involves various indications of the multiplicity, number, quantity, or amount of some object of noun phrase relative to a predication (23). Formally, it includes such things as plural inflection, plural concord, and indication of singular or plural "by modification of the lexical head of a noun phrase with a numeral or adjective indicating quantity or specificity of amount" (23). 

Both English and Yucatec represent number in the noun phrase and do so in somewhat similar ways: they both mark plural by inflection and concord; they both mark singular by adjectival modification (56). However, the overt expression of plural is more important in English, where it is obligatory for a large range of noun phrase, than it is in Yucatec, where it is optional for a relatively small range of noun phrases (56). 

The difference in typical patterns of pluralization in English and Yucatec can be characterized more precisely in groups of lexical noun phrases. (See table 2.) Group 1 noun phrases refer to animate beings and similar self-segmenting entities (e.g., dogs, automobiles); Group 2 noun phrases tend to refer to discrete objects and similar stably segmented entities (e.g., shovels); and Group 3 noun phrases tend to refer to tangible materials or substances with malleable form and similar segmentable entities (e.g., mud) (58). 

Table 2: Scope of typical pluralization in English and Yucatec lexical noun phrases in terms of preliminary formal groupings (Lucy 58). 

Noun phrases marked [+ animate] specifically signal reference to an animate entity. By contrast those marked [- animate] are not specified one way or the other as to animacy from a linguistic point of view. However, this group is subdivided into discrete (distinct, stably bounded objects) and non-discrete (materials or substances).

Picture Practice

Twelve Yucatec men and twelve US men were asked to participate in Lucy’s picture task series. (See Figures 2-7.) The first task required speakers to describe verbally what they saw in figure 2, allowing the respondent to become familiar with the stimuli in the context of a readily understandable procedure. The second task followed the pattern of the first, but was more difficult in that it required the verbal report to be done without the picture in view—that is, on the basis of short-term memory. The third task involved presenting the respondent with figures 3-7 and asking him to pick which of the pictures were most similar to the original. The fourth task involved short-term recognition memory in which each individual was asked to select the original he had seen in task 1 and 2 from all six figures. All six figures differ in slight ways and test variations in quantity in all three groups of lexical noun phrases. (See table 3.)
 

Figure	Differentiation
1	original
2	boy missing
3	bottle missing
4	broom added
5	extra corn by hen
6	extra corn by pig

 

Table 3: Inventory of standard objects and target objects by picture (Lucy 166). 

Results 

English speakers indicate number for almost every animal and non-animate discrete object (Figures 2-4). They indicate numbers of substance-type objects significantly less frequently with conformity to table 2. By contrast, Yucatec speakers indicate number much less often overall. Further, within this lower level of specification, Yucatec speakers indicate number frequently for animals and significantly less frequently for non-animate objects (both discrete and nondiscrete) (104).

Based on these results, Lucy concludes that English speakers habitually attend to the number of various objects of reference more than Yucatec speakers (87). Differences in linguistic structure "in the form of regular patterns of morphosyntactic markings index clear differences in informational content with respect to number as a function of object type in these two language" (106). In using plural less frequently, the Yucatec speakers consequently view picture scenes differently and notice less variations opposed to the English speakers. The frequency of pluralization in each language influences both the verbal and nonverbal interpretation of pictures (157). 

Language, therefore, affects a speaker’s habitual dispositions toward, or ways of responding to, a stimuli. The very means by which we communicate can affect our interpretation, memory, manipulation, and decision (91). 

DISCUSSION

These studies have certainly resuscitated a dying linguistic hypothesis. By dovetailing Whorf’s theory in 1956, Lucy and Slobin have demonstrated that language can directly influence our thoughts. Through verbal limitation, grammatical focus, and structural emphasis, oral communication can pattern our very way of thinking. Cultural anthropologist Andy Clark concludes that language not only "confers on us added powers of communication; it also enables us to reshape a variety of difficult but important tasks into formats suited to the basic computational capacities of the human brain" (194). Hence, cultures with different structural axioms result in different computational capacities. 

Yet, a dangerous consequence results in cultures claiming cognitive acuity based on their language structures. Even though American English includes an additional lexical noun phrase group compared to the Yucatec Mayan, cognitive dominance should not be attributed to the English. Language structures merely influence another way of thinking, not necessarily a better one. In the predominantly rural Yucatec Mayan culture, for example, little importance is attached to non-animate, discrete objects (See table 2). Since this culture does not need to differentiate between quantities within this lexical noun group, their lack of distinction is linguistically conservative. In English-speaking areas, however, it is often necessary to differentiate among items in this group. Therefore, language differences do affect thought but should be viewed in light of a cultural setting.

The Whorfian hypothesis can perhaps shed light upon cultures, themselves. Variations in thoughts and languages can ultimately pinpoint cultural differences. These external outputs can thus provide valuable clues for implicit community laws, religious beliefs, and unrecorded customs. By studying the variations in language, we can understand variations in thoughts and place them in the context of a cultural setting. In doing so, cultural linguists gain an even greater insight into the societies.

Such a task, unfortunately, seems overwhelmingly difficult. While research like that of Slobin and Lucy provide beginning data, only small portions of languages have yet to be studied. In addition, only a few cultures have even been compared. As a result, an exceedingly large domain of research remains. Until further studies have been conducted using the Whorfian hypothesis, conclusive universal statements cannot be drawn among different cultures.

Perhaps, then, a more appropriate place to begin research is internally—within the individual, within a language, and then between cultures. Our own languages are so complex and variable that, until more accurately understood, cannot be compared cross-culturally. How can two entities with unknown qualities be compared? To compare two languages, the linguist must understand both languages. Research done by Kay, Bivens, and Berk all suggest that the Whorfian hypothesis can apply on a personal and intra-linguistic scale. By studying which language structures influence our own thoughts, we can then compare these findings to that of other cultures. Interesting and insightful, such research is also painstakingly complex and time-consuming.

As neuroscientific technology improves our clinical interpretation of language, linguistic studies such as the Whorfian hypothesis will advance our cultural conception within and between languages. The extent to which language ultimately affects thought may forever remain unknown, but language’s impact will be anything but sketchy. The theory of linguistic relativity, after period of dormancy and rapid decline, may soon become a cultural doctrine. 

Brown, R and E. Lenneberg. "A Study in Language and Cognition." Rethinking Linguistic Relativity. Eds. John J. Gumperz and Stephen C. Levinson. Cambridge: Cambridge University Press, 1996.

Clark, Andy. Being There: Putting Brain, Body, and World Together Again. Cambridge: The MIT Press, 1997.

Gumperz, John J. and Stephen C. Levinson. "Introduction to Part I." Rethinking LinguisticRelativity. Eds. John J. Gumperz and Stephen C. Levinson. Cambridge: Cambridge University Press, 1996.

Kay, Paul. "Intra-Speaker Relativity." Rethinking Linguistic Relativity. Eds. John J. Gumperz and Stephen C. Levinson. Cambridge: Cambridge University Press, 1996.

Lucy, John A. Grammatical categories and cognition: a case study of the linguistic relativity hypothesis. Cambridge: Cambridge University Press, 1992.

——. "The Scope of Linguistic Relativity: An Analysis and Review of Empirical Research." Rethinking Linguistic Relativity. Eds. John J. Gumperz and Stephen C. Levinson. Cambridge: Cambridge University Press, 1996.

Pinker, S. Learnability and cognition: the acquisition of argument structure. Cambridge: MIT Press, 1989.

Slobin, Dan. "From ‘Thought and Language’ to ‘Thinking and Speaking.’" Rethinking Linguistic Relativity. Eds. John J. Gumperz and Stephen C. Levinson. Cambridge: Cambridge University Press, 1996.

Whorf, Benjamin. Language, thought, and reality: selected writings of Benjamin Lee Whorf.  Ed. J.B. Carroll. Cambridge: MIT Press, 1956.