Title: A discussion and extension of Davidoff 2001' Language and Perceptual Categorisation
1A discussion and extension of Davidoff (2001).
Language and Perceptual Categorisation
- Kelly Sorensen
- Christopher Thomas
- November 2, 2004
2Outline
- Jules Davidoff
- The Sapir-Whorf Hypothesis
- Discussion of Language and Perceptual
Categorisation - Discussion of Linguistic Relativism
3Prof. Jules Davidoff
- Professor of Psychology, Goldsmiths College,
University of London - Research
- mental representation of objects
- relationship between the stored (memory)
knowledge concerning objects and their
recognition, categorisation and nameability - effects on the way speakers of a language
perceive, categorize and remember colors
4Outline
- Jules Davidoff
- The Sapir-Whorf Hypothesis
- Discussion of Language and Perceptual
Categorisation - Discussion of Linguistic Relativism
5Sapir-Whorf examples
- Eskimos have four different words for snow, where
English has just one - aput for snow on the ground
- qana for falling snow
- piqsirpoq for drifting snow
- qimuqsuq for a snowdrift
6Whorfs conclusion
- "We have the same word for falling snow, snow on
the ground, snow packed hard like ice, slushy
snow, wind-driven flying snow -- whatever the
situation may be. To an Eskimo, this
all-inclusive word would be almost unthinkable
he would say that falling snow, slushy snow, and
so on, are sensuously and operationally
different, different things to contend with he
uses different words for them and for other kinds
of snow."
7Whorfs conclusion
difference in attitude or perception
difference in vocabulary
8Introduction to the Sapir-Whorf Hypothesis
- In linguistics, the Sapir-Whorf Hypothesis states
that there are certain thoughts of an individual
in one language that cannot be understood by
those who live in another language. - The hypothesis states that the way people think
is strongly affected by their native languages. - It is a controversial theory championed by
linguist Edward Sapir and his student Benjamin
Whorf.
9Historical Notes
- Whorf was fighting against cultural evolutionary
theory saying that Western thought is the highest
form of thought - Sapir and Whorf rejected hierarchical,
quasi-evolutionary rankings of languages and
cultures .in particular the European, especially
Humboldtian, obsession with the superior value of
inflectional languages for the cultural or mental
advancement of a people. (Lucy 1997)
10Historical Notes
- After vigorous attack from followers of Noam
Chomsky in the following decades, the hypothesis
is now believed by most linguists only in the
weak sense that language can have some small
effect on thought.
11Sapir-Whorf hypothesis I
- Linguistic relativity
- Structural differences between languages are
paralleled by nonlinguistic cognitive differences
(the structure of the language itself effects
cognition) - The number and the type of the basic colour words
of a language determine how a subject sees the
rain bow
12Sapir-Whorf hypothesis II
- Linguistic determinism extreme "Weltanschauung
version of the hypothesis - The structure of a language can strongly
influence or determine someones World View - A World View describes a (hopefully) consistent
and integral sense of existence and provides a
theoretical framework for generating, sustaining
and applying knowledge - The Inuit can think more intelligently about snow
because their language contains more
sophisticated and subtle words distinguishing
various forms of it, etc.
13Sapir-Whorf hypothesis III
- Arbitrariness
- The semantic systems of different languages vary
without constraint. - This hypothesis must be tacitly assumed, because
otherwise the claim that Linguistic Relativity
makes is rather undramatic. - For each decomposition of the spectrum of the
rain bow a natural system of colour words is
possible
14Outline
- Jules Davidoff
- The Sapir-Whorf Hypothesis
- Discussion of Language and Perceptual
Categorisation - Discussion of Linguistic Relativism
15Why/how do we determine category membership?
- Davidoff (2001) argues
- that it is linguistic similarity rather than
perpetual similarity that is critical for
perceptual categorization - against the view that there are underlying,
universal, neurophysiological mechanisms which
determine how color is categorized
16The case for universal color categories
- Is based on knowledge of how wavelength-sensitive
neurons function. - Based on the opponent-process mechanism of
neurons, it has been argued that there are two
elemental achromatic categories - Black
- White
- and four elemental color categories
- Red
- Green
- Yellow
- Blue
17The case for universal color categories cont.
- There are two wavelengths for which
opponent-process neurons R-G give no output. - There is also a wavelength for which the
opponent-process neurons Y-B give no output.
18Problems with the case for universal color
categories
- Wavelengths chosen to represent the colors blue,
yellow, and green are not consistent with what is
expected based on neurophysiology - Conclusions about neurons are weakened by
individuals previous exposure to the concept of
blue, yellow, or green
19Problems with the case for universal color
categories
- Neurophysiology data show that neurons can
respond selectively to particular wavelengths or
combinations of wavelengths and brightness no
evidence, however, that neurons respond
categorically. - Davidoff thus concludes that perceptual
categories cannot be based strictly upon
observation.
20The philosophical argument
- The Sorites paradox
- Take a series of colors of decreasing wavelength
with the change below the threshold for the human
visual system - Agree that a patch at one end can be called red
- If red is the observational or perceptual
category, then the next patch must also be called
red, and so on. - Continuing with the logic we come to the
illogical conclusion that all colors in the
series are red, even the blues at the other end
of the series
21(No Transcript)
22(No Transcript)
23(No Transcript)
24(No Transcript)
25The neuropsycological evidence
- Patients with language impairments caused by
brain damage often behave as if the Sorites
paradox is a reality, sorting by perceptual
similarity without regard for categorical
boundaries.
26Cross-cultural theories
- Whorfian view-We dissect nature along the lines
laid down by our language. - Rosch- argued for a universal rather than
language based color categories due to cognitive
similarities between languages with few color
terms and English.
27Roschs universal theory
- Based upon studies of the Dani who
- have only two basic color terms but remembered
colors in ways very similar to the English
speakers - Showed superior learning and memory for focal
colors for which they had no linguistic terms - Results were widely accepted as proof of
universal color categories
28Problems with Roschs universal theory
- Davidoff argues that there are potentially
serious flaws in both the design and
interpretation of Roschs studies - Conflicting results found for the first study on
two measures based on the multi-dimensional
scaling of the same data - Graphical demonstration showed support for
universalist view - Statistical results did not
29Problems with Roschs universal theory cont.
- No explanation is given for conflicting results
- Dani speakers perform poorly on the statistical
measures for subsequent studies as well - Researchers unable to replicate findings with
Berinmo population from New Guinea
30Categorical perception
- Stimuli from the center of perceptual categories
are classified faster than those at the edges,
consequently discrimination of stimuli is better
across than within categories - In studies with Berinmo and English speakers,
classification was consistently more closely
aligned with the linguistic categories than with
the underlying perceptual universals
31Empirical support for Whorfian view (theory of
linguistic similarity)
- 1st experiment
- When making judgements of similarity between a
group of three stimuli, participants judged two
stimuli from the same linguistic category to be
more similar, even thought the perceptual
distance between each pair of stimuli were held
equal - No reliable tendencies were observed for those
belonging to groups which make no linguistic
distinctions between the categories used -
32Empirical support for Whorfian view (theory of
linguistic similarity) cont.
- 2nd experiment
- English speakers
- found the division between green and blue easier
to learn than the arbitrary division of green - found the division between yellow and green
easier to learn than the division between the
Berinmo color categories of nol and wor
33Empirical support for Whorfian view (theory of
linguistic similarity) cont.
- Berinmo speakers
- Demonstrated no difference in difficulty for
learning the green-blue division and the
arbitraty green division - Found the nol-wor division significantly easier
to learn than the yellow-green division
34Empirical support for Whorfian view (theory of
linguistic similarity) cont.
- 3rd experiment
- Demonstrated an effect of linguistic category in
recognition memory - English speakers showed significantly superior
recognition for targets from cross-category pairs
than for those from within-category pairs for the
green blue boundary, but not for the nol-wor
boundary - 2. Berinmo speakers showed the opposite effect
35English and Berinmo Color Categories
English color categories
Berinmo color categories
comparison
36Davidoffs conclusions from these 3 experiments
- Categorical perception shows the influence of
language on perception - The structure of linguistic categories distorts
perception by stretching perceptual distances at
category boundaries
37Interference studies
- Has examined whether categorical perception can
be disrupted with verbal interference - Verbal interference removed the cross-category
advantage for speakers whose languages classifies
the colors as belonging to different categories - It appears that verbal coding (representation of
information verbal) facilitates recall
(information is likely encoded both visually as
well as verbally)
38Constraints on Whorfian view
- The argument for color categories being a product
of language does not mean that categorization is
unrelated to properties of the visual system - Similar items (as defined by perceptual
discrimination) are universally grouped together
(e.g. would not have yellow and blue together
without also having green between) - Even perceptual categorization tasks can
sometimes be solved simply by perceptual
similarity or common association
39Overall conclusions of the author
- Perceptual categorization is determined by
linguistic relativity - Being able to attend to color is different from
understanding color categories
40Overall conclusions of the author
- Cross-lingual evidence supports the Whorfian
hypothesis in the number domain, in space, in
time, and in speech perception - Language and cognition interact children
generalize abstract terms only if they have
learned a label for the concrete-learning
situation
41Questions for future research
- Can human-primates form perceptual categories?
- There is evidence that neonates show color
categorization. Does this reflect categorization
of a different type? - Are there capacity constraints on perceptual
categorization?
42Questions for future research
- Verbal interference affects categorization in
memory tasks. Is the same true for perceptual
tasks? - Which brain areas are involved in
perceptual-categorization?
43Outline
- Jules Davidoff
- The Sapir-Whorf Hypothesis
- Discussion of Language and Perceptual
Categorisation - Discussion of Linguistic Relativism
44More Evidence in favor of linguistic relativism
- Chinese children count earlier than American
children - (In part) because Chinese numbers are more
systematic
45More Evidence in favor of linguistic relativism
Mayans similarity judgments are more influenced
by material (as appropriate for mass nouns),
rather than shape (as appropriate for count nouns)
46But then
- What does this evidence really say about the
influence of language on thought? - Especially in the case of colors, is it more a
matter of what weve learned? - Painters can name more colors.
- We can look at colors from different points of
view - Warm or cold colors
- Pastel or vivid colors
47Pinker against Sapir-Whorf
- Supposed limitations on expression in various
languages are based on faulty linguistic
understanding. - Hopi does have words for time, etc.
- Translation between languages is possible (even
if difficult to do elegantly).
http//www.ling.upenn.edu/courses/Spring_2002/ling
001/thought.html
48Pinker against Sapir-Whorf
- Thought is possible without language.
- Adults who have grown up without language.
- Babies before they learn language.
- Primates and other animals that never learn
language. - Adults who reason and create in visual or other
modes.
http//www.ling.upenn.edu/courses/Spring_2002/ling
001/thought.html
49Pinker against Sapir-Whorf
- Language is an inadequate medium for the direct
encoding of thought. - We often can't think of the right word to express
ourselves. - Language contains ambiguity, homophony, etc.
- Manipulation of visual images is done directly.
- ? Pinker suggests a nonverbal language which he
calls Mentalese
http//www.ling.upenn.edu/courses/Spring_2002/ling
001/thought.html
50Questions
- Do euphemisms make us think differently about a
fact? - negative growth, collateral damage, peace force
- retarded, mentally disabled, mentally challenged
- Does the convention of using the male form
support patriarchic views or is it just an
indication?
51Questions
- Can we think about categorization without
language? - For visually similar items
- For abstract categorizations
52The common denominator
- In an experiment similar to Davidoffs, Kay and
Kempton came to the following conclusions - The extreme ("Weltanschauung") version of this
idea, that all thought is constrained by
language, has been disproved - The opposite extreme that language does not
influence thought at all is also widely
considered to be false
53Synthesis?
- Looking at Pinkers objections, is there
something underlying language that is more
influential? - Is it the language that influences our perception
or rather the culture we live in? - Lakoff suggests that cultures have deeply rooted
conceptual metaphors that - find their expressions in the language
- guide our perception
54References
- Paul Kay, Willett Kempton What is the
Sapir-Whorf Hypothesis? American Anthropologist,
New Series, Vol. 86, No.1, March 1984, 65-79 - http//www.ling.upenn.edu/courses/Spring_2002/ling
001/thought.html - Jules Davidoff Language and Perceptual
Categorisation. TRENDS in Cognitive Sciences
Vol.5 No.9 September 2001, 382-387 - George Lakoff Mark Johnson. (1980) Metaphors We
Live By. Chicago University of Chicago Press. - John A. Lucy. Linguistic Relativity. Annu. Rev.
Anthropol. 1997. 26291-312