Sapir & Whorf

The Sapir-Whorf Hypothesis


Why does it matter that language may be sexist, racist and so on?  Apart from the obvious point that such language is offensive, many people that the language we hear and use has a big influence on the way we think.

There are two extreme views of this issue, the universalist and the relativist.

The universalist position is that all humans share common ways of thinking, a set of basic concepts about the world which we may call conceptual primes.  One example is relative distance, the distinction between “near” and “far”.  All languages, whatever their apparent differences, will provide means of expressing these essential concepts.  According to this view, language simply reflects our thoughts.  For example, racist terms exist because people have racist attitudes.  The notion that language reflects thought is known as reflectionism.

The relativist position is the opposite of the universalist.  We rely on language to form our ideas.  Individual languages differ greatly in both lexis and grammar.  It follows that the speakers of different languages will experience and understand the world in very different ways.

This position is mainly associated with two American linguists, Edward Sapir (1884-1939) and Benjamin Lee Whorf (1897-1941).  The Sapir-Whorf hypothesis as it is usually known, is that language actually determines thought.  This theory is described as linguistic determinism

Whorf studied the language of the native American Hopi people and observed that it was “timeless“.  It lacked time adverbs and did not mark verbs for tense as we do in English.  Hopi grammar was more concerned with distinguishing between what is objective and what is subjective.  Whorf concluded that a Hopi speaker must view the world very differently from a native English speaker. 

Other studies have focused on colour words.  Some languages have more words than others for labelling colours.  Many have around a dozen basic colour terms while others have as few as four. Different languages use colour words to divide up the spectrum in different ways.  Research into the Mexican language Tarahumara, which has only one word to cover both blue and green, suggested that its speakers distinguished between these colours less well than English speakers.

More recent work has largely discredited this extreme view, however. 


  • Even if we do not have a word or structure equivalent to one in another language, we can still find ways of expressing an idea from that language, using the words at our disposal.  For example, Australian aborigines whose native languages lack words for numbers can still learn numbers and be taught arithmetic.
  • It is doubtful whether we actually think in words.  How often have you heard someone say, “I know what I mean but I can’t put it into words”?  Some famous thinkers have claimed to rely more on mental images than language.
  • Language changes.  New words appear regularly.  Could this happen if our thinking was limited by our existing language?


The Sapir-Whorf hypothesis, then, is no longer accepted in its absolute form.  Nevertheless, there is plenty of evidence that our thinking is influenced by language.  This can be seen as the weak version of the hypothesis

Elizabeth Loftus investigated eye witness testimony and its reliability.  How easy is it to influence a witness’s recall of an event?  She showed her subjects videos of traffic accidents and asked them to fill in a questionnaire.  However, there were different versions of the questions.  Some people were asked to estimate the speed of the vehicles “when they contacted”.  For other groups, the verb “contacted” was replaced by “hit”, “bumped”, “collided” or “smashed”.  The answers she received differed noticeably.  The more violent the verb used, the higher the estimated speed.  A week later, she asked some of her witnesses whether they had seen any broken glass.  (There had been none).  Again, the people who had been asked about cars “smashing” into each other were far more likely to say yes than those who had been asked about cars “hitting” each other.

You have probably studied persuasive language.  Rhetorical techniques such as the “power of three” are used because we see that they work. 

 However, all these examples show people being influenced by the language used to represent ideas or events to them.  This is not the same thing as saying that our native language structures our ability to think.

 Most linguists today accept that language and thought are interdependent.  Sexist language exists because of sexist attitudes.  But growing up in a society where such terms are current may encourage people to accept or adopt those attitudes.


            An important debate in linguistic theory is whether thought controls language or language controls thought. In their most extreme forms, the first of these positions is called reflectionism, the second determinism.

            The Sapir-Whorf Hypothesis suggests that language divides the complex flux of impressions we experience from the world into arbitrary categories. The result of this is that the way we perceive the world is influenced by the way our language organises and classifies it.

            However, the question as to just how much impact this has on thought has led to different versions of the hypothesis developing. Moderate Whorfianism asserts that thought processes are merely influenced by language organisation, whilst extreme Whorfianism says that thought processes are completely controlled by it. Therefore, in this latter form, the hypothesis is deterministic – i.e. the language you speak determines how you can think, or that you cannot have a certain thought without a particular word for it.

(However, Sapir and Whorf distanced themselves from the behaviourist stance that there is no difference between language and thought, thought being entirely linguistic.)


Linguistic relativism – the gulf between different languages


            Whorf, who was a student of Sapir, brought the hypothesis close to linguistic determinism. He developed the related idea of linguistic relativity: this suggests that, due to the control language has over thought, people who speak different languages perceive and think about the world quite differently from each other. He gave the example of Inuits having a supposedly large number of words for snow, and suggested that this meant they could better perceive minute differences in weather conditions than people of other languages.

            Analysis: There is some support for this. Researchers Kay & Kempton tested this   by asking participants to look at different coloured chips. They found that people      who possessed more specific colour words in their vocabulary were better able to            perceive differences in colour, ie, if you have a word for light-blue, your chances            of perceiving a difference between a light-blue and dark-blue chip may be enhanced. However, Pinker has dismissed this experiment as flawed, suggesting that all it showed was that subjects remembered of each chip both a non-verbal            image and a verbal label, as two forms of memory would be more reliable.

            This means that it can be difficult to translate language. Translating occurs between languages but also, Whorf argues, by reformulating within the same language the words used to express a thought. If you change the words with which you express an idea, you aren’t just expressing the same idea ‘in a different way’; you are expressing a subtly different idea. Further, this suggests that meaning does not reside ‘in’ a text itself, particularly in expressive forms of communication such as poetry, but rather is brought to the text through readers’ responses to specific words, which is in turn shaped by the reader’s cultural context.

            Analysis: There is both research and anecdotal evidence that ‘translating’ ideas      between different words can be problematic. An experiment by Loftus aimed to         gauge the impact of word usage on memory, specifically eyewitness testimony.             Showing participants the same film of a car crash, she then used different words –             contacted, hit, collided, smashed into – to trigger their recall. She found that this   affected how they remembered the film, suggesting that language used does    influence thought. Further, many authors and poets have resented the translation           of their works into different languages, feeling that changing the words used to             communicate their meaning will drastically alter the meaning itself. However,        others suggest that ‘no language is untranslatable’ as basic ideas (often scientific)    remain the same no matter how they are verbally expressed – this is universalism.         A cross-cultural version of Kay & Kempton’s study by the BBC found that people from 100 different societies grouped the 300 chips into the same six basic colour   groups.

The hypothesis today


            Although few support the hypothesis in its deterministic form, many linguists now accept the ideas of moderate Whorfianism. As well as emphasising the influence rather than control language has over thought, they suggest that it is a two-way process and the way we see the world, as communities or individuals, affects the language we use; and stress the importance of social context in shaping language use. However, some dismiss the hypothesis completely in favour of universalism. Indeed, evidence such as the BBC’s study and the Inuits’ fictional words for snow does seem to suggest that words used for specific concepts (especially if they are visibly observable) do not fundamentally change people’s perceptions of them. Nevertheless, this does not disprove the Sapir-Wharf hypothesis by any means. It seems likely not only that the verbal presentation of abstract concepts should do much more to affect their meaning than it does to concrete concepts such as colour, but also that language in the way it is formed, eg, questions, negation, line of argument leading to a conclusion, etc, affects more fundamentally the way that humans think.

Ways of testing the Sapir-Whorf hypothesis


            If our thoughts were really influenced by the language at our disposal, it follows that change in the terms a language uses for certain things will also alter people’s ideas of the things themselves. A good example of this is the growth of political correctness in the UK.

            Political correctness is a movement that started in the 1970s that aimed to change the language used towards minority groups and women, thus behaving more politely towards these groups. Established ways of referring to these groups in ways that often showed low regard for them were challenged and began to change. It shows the extent to which political correctness has made an impact that many of such commonplace terms from the 1970s are regarded as highly offensive today.

            However, the prejudice remaining in today’s society means that the actual effectiveness of political correctness is unclear. Reflectionists would argue that this is because the human impulse to fight for superiority by belittling others is still constant, regardless of whether it is expressed differently today. But there is the possibility that the development of less offensive terms for women and minority groups has lessened people’s prejudice against them.

            I am going to test this and thus the Sapir-Whorf hypothesis. I shall first establish whether certain words create a more negative picture of women than similar ones do of men, and then see whether changes in terms used to describe women has reduced the negative response.


a) Terms referring to a certain concept in women will provoke more negative connotations than terms describing a similar concept in men (see below, ‘gawky’ v.s ‘dowdy’, ‘slut’ v.s ‘rake’).

b) Terms used for women during the 1970s (‘housewife’, ‘actress’) will provoke more negative connotations than modern terms (‘stay-at-home mum’, the gender-neutral ‘actor’). 

METHODOLOGY: I presented 3 people with the following instructions and list of words, and gave them five minutes to record their responses.

What do these words make you think of?

What kind of person are these words describing?

What would they be like? How would they behave?

Housewife, Actor, Slut, Gawky, Actress, Stay-at-home mum, Dowdy, Rake

RESULTS: Housewife

Middle aged  working class woman, during 1940’s or 1950’s- caring little for her appearance or her own needs, doing tedious tasks for others. Proaly aproned and wearing scarf over her hair

Dowdy, small, unambitious, unassuming, old fashioned as a social role

Downtrodden woman, old-fashioned, dull, ignored, cleaning/cooking all day


Flamboyant quirkily dressed person with clear anunciation and distinct mannerisms, male, youngish

Talkative, extravert, demonstrative, flamboyant, observing others’ behaviour, mannersims and speech

Loud, self-confident, charismatic, cultured


Female young tending to plump, little care in her appearance,  for example laddered tights grubby clothing  too tight and slightly too small, underwear visible, promiscuous, working class, little education

Loud, blonde, make up, swaggering, overtly sexual in speech and behviour

Female, promiscuous, “lower-class”, cheap, wears lots of make-up but not very well, emotionally shallow



Tall awkward looking quality of a person, usually male

Clumsy, shy, male, adolescent, long limbed, all fingers and thumbs

Awkward mannerisms, male, socially unskilled, teenager, long-limbed


Female more demonstrative and extroverted than the male version but otherwise fairly similar

Insightful, articulate, stylish, ambitious, highly socialised with other men and women

Diva, flighty, glamourous, vain, melodramatic, egotistical, not very intelligent

Stay-at-home mum

Modern woman who has made positive choice to look after her children

Hardworking, conscientious, gossippy, maternal, would love a proper job but has sacrificed this

Caring, has made a choice not to work to help her children



Frumpish woman, lazy, dull

Female, small, hunched, grey, old fashioned with limited horizons

Older woman, small-minded, dresses drably, unattrative – has let herself go

Rake (not the garden implement kind)

Young male, historical figure tending to hedonism, misogynistic no doubt

Male, young, rich, languid, upper class young man of the old order

Male, young, rich, allowed licence to behave how he wants, successful with women but won’t commit


a) Responses also differed in the terms applying similar concepts to males and to females. For example –

Common ideas about ‘slut’/number of times occurring

Too sexually active – 5

Appearing seedy/unattractive – 3

Not very intelligent –2

Common ideas about ‘rake’/number of times occurring

Sexually successful/indulgent – 2

Wealthy/upper class – 3

Allowed freedom – 2

In contrast to the former response, the only implication ‘rake’ indicating sexual excess is in the description ‘tending to hedonism’ – and even this more suggests a chosen way of life rather than a character flaw. Meanwhile, the word ‘gawky’ was seen as more a physical trait, rather than a sign of unattractiveness or a flawed character as ‘dowdy’ was.

b) There was significant difference in response to 1970s terms compared to modern terms. For example –


Common ideas about ‘housewife’/number of times occurring

Low status – 3

Unimportance as an individual – 6

Old-fashioned – 3

Common ideas about ‘stay-at-home mum’/number of times occuring

Choice/ability to do other things – 4

Positive feminine attributes, nurturing instinct etc – 3

The difference in attributed qualities suggests that the participants’ ideas about the concept were shaped by the words they used. The responses to ‘actor’ and ‘actress’ were also different, with ‘actress’ seen as more social, but the difference was less marked. On reflection, this is probably somewhat due to ambiguity in the test material – at least one participant assumed ‘actor’ referred exclusively to males, rather than being an inclusive gender-neutral term as I intended it.


a) As comparing terms for similar male and female attributes showed, language carries prejudices within it against certain groups, here, women. This conclusion alone could be used to support either reflectionism or determinism. However –

b) Comparing 1970s terms for women with modern terms for women suggests that it is often the words themselves that carry the prejudice. Changing terms used for minority groups and women can actually change people’s views about them, as political correctness has done. This suggests that language at least influences the way we think, supporting the Sapir-Whorf hypothesis.

1970: Your wife won’t be able to stall the car or grind the gears – and there’s even one pedal fewer to confuse her with the Mini Automatic



R Lewis


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s