Niemeier, Susanne and Ren� Dirven, ed. (2000) Evidence for
Linguistic Relativity, John Benjamins, hardback, xxi, 240
pp., Amsterdam Studies in the Theory and History of
Linguistic Science 198.
S�ren Wichmann, Danish Institute for Advanced Studies in the
Humanities
The last decade or so has seen an upsurge in the interest in
the so-called Sapir-Whorf hypothesis. The idea that languages
excert an influence on the thought and behavior of their
speakers has fascinated many linguists, but few would claim
that the hypothesis is easy to test, and some might even
consider it beyond the reach of empirical approach. Thus it
is interesting to see the publication of a book that promises
actual evidence for linguistic relativity. Most of the papers
in this edited volume were presented at a symposium on
"Humboldt and Whorf Revisited: Universal and Culture-Specific
Conceptualizations in Grammar and Lexis", held at Gerhard-
Mercator University, Duisburg, 1998. Another group of papers
from that meeting is published in the same series in a volume
entitled Explorations in Linguistic Relativity (see LINGUIST
11.1025 for announcement with short descriptions of both
volumes.) The book is structured in two parts, "Part 1:
Evidence from Language: Production, Interpretation, and
Change" (pp. 1-103) and "Evidence beyond Language: Cognition,
Discourse, and Culture" (pp. 107-233). Since, in the mind of
the present reviewer, this division is in some of the cases
arbitrary the contributions are reviewed in a slightly
scrambled order in order to highlight links between some of
them.
The volume opens with John Lucy's "Introductory
Comments" (pp. ix-xxi), that situates the papers within broad
a typology of ways to approach linguistic relativity.
The paper "linguistic Relativity in Speech Perception:
An Overview of the Influence of Language Experience on the
Perception of Speech Sounds from Infancy to Adulthood" (pp.
1-28) by Ocke-Schwen Bohn begins with the following
statement: "For a speech scientist, linguistic relativity is
not a hypothesis, it's a fact." In his well informed essay
Bohn guides us into a universe that is not often considered
by students of linguistic relativity, showing that speech
perception from the child's very first stages of language
learning is highly influenced by the patterns of the first
language (L1). It is not obvious that semantic categories
necessarily determine our perception of reality in the same
way that our habitual attention to phonological patterns of a
L1 influences our perceptions of the sound patterns of other
languages, however, so it is questionable whether studies of
speech perception really provide evidence for the Sapir-Whorf
hypothesis as such. Even if Bohn's study only provides an
analogical argument from a different area--that of task-
specific behavior--it is nevertheless relevant and thought-
provoking since it describes models for rigorous
experimentation and data handling that defenders of
linguistic relativity should perhaps live up to were they
truly to prove their hypothesis.
It is interesting to read the paper "Can Grammar Make
You Feel Different?" (pp. 53-70) by Michael Maratsos, Demetra
Katis, and Annalisa Margheri in conjunction with Bohn's
paper, since Maratsos et al. actually make an attempt to test
Whorfian effects by devicing a research design that involves
rigorous experiments open to statistical evaluation--and
fail. The three authors present the results of experiments
testing whether the different grammars of the verbs 'like'
and 'miss' in Italian and Greek as opposed to English produce
different ratings of control assigned to the Experiencer by
speakers of the three languages. According to a hypothesis
due to Izchak Schlesinger predicate roles will tend to take
on a grammatical coloring of the grammatical roles to which
they are assigned, grammatical subjects being somehow more
agentive than direct objects. Thus, in Italian and Greek
translation equivalents of 'John likes Mary' and 'John misses
Mary', where the Experiencer (John) is expressed as an
Oblique Object and the Stimulus (Mary) is the grammatical
Subject, the Experiencer would be expected to be assigned
less control than in the corresponding English sentences,
where the Experiencer is Subject and the Stimulus Object.
This was tested by asking native-speaker subjects to mark off
degrees of control for respectively Experiencer and Stimulus
in sentences involving different kinds of Experiencer verbs
on a seven-point scale. Interestingly, there turned out to be
no significant differences in the ratings of 'like' and
'miss' as opposed to other Eperiencer verbs from Greek and
Italian speakers on the one hand to English speakers on the
other. The authors conclude that it is not given that their
experiment disproves the existence of Whorfian effects and
mention as one possibility that Whorfian effects are more
likely to be produced by more overall grammatical patterns
than just the peculiar grammars of single items like the two
verbs studied. Another conclusion, which seems to be more
immediate, although the authors for one reason or the other
do not stress it, is that the theory that grammatical
Subjects are more agentive and Patients more patientive is
simply wrong. In any case the study is not very conclusive
since there seem to be a large number of possible cultural
factors as well as factors relating to the research design
itself that could conceivably have influenced the results.
Minimally the study shows that it is not an easy task to
prove the existence of Whorfian effects experimentally.
Although it is difficult it does not seem to be
impossible to demonstrate Whorfian effects by direct
experimentation, as the paper "Universal Ontological
Knowledge and a Bias toward Language-Specific Categories in
the Construal of Individuation" by Mutsumi Imai (pp. 139-160)
demonstrates. Focusing on the domain of individuation and
contrasting data from English and Japanese speakers Imai sets
out to ask whether different languages influence the
construal of objects and substances as individuated or non-
individuated and whether children aquire ontological concepts
independently of or through language. Her conclusion is that
some ontological knowledge is language-independent although
there is also a significant contribution to the construal of
ontological categories from language. For instance, from as
early as the age of two children, independent of their
language, will tend to treat substances as non-individuated
and objects as individuated, by sorting substances according
to their material and objects according to their shape. Over-
all the tendency to categorize both objects and substances
according to shape, treating both as individuated, is
markedly greater with Americans than with Japanese once the
subjects reach 4 years of age. Through experiments of this
kind Imai actually makes a plausible case for Whorfian
effects.
G�bor Gy�ri's argument in his essay "Semantic Change as
Linguistic Interpretation of the World" (pp. 71-89) is that
semantic change is prompted by the need to express new ideas
and represents a transformation of linguistic expressions for
familiar experience to forms that capture the new ideas.
Since, according to Gy�ri, the linguistic expressions from
which the transformations depart set certain limits on the
outcome, semantic change proves the validity of the
linguistic relativity hypothesis. Occasional examples taken
from The American Heritage Dictionary of Indo-European Roots
are called upon to support this argument. Gy�ri makes no
empirical inquiries into the actual circumstances of semantic
change by studying incipient changes, but simply assumes that
the retrospective descriptions of semantic changes in terms
of transfers such as metaphor, metonymy etc. correspond to
actual cognitive operations in the speaker's minds and serve
as explanations for the way semantic change works. In the
mind of the present reviewer it is difficult to accept that
the point of departure of the change as well as the endpoint
should simultaneously be available as stations for cognitive
operations in the speakers, and it is also difficult to
accept that language-specific restraints on the semantic
material available for modification for the purpose of
conceptualizing new ideas should provide evidence for
linguistic relativity. One of the examples cited by Gy�ri's
is that of Eng. wheel from PIE *kwel- 'to revolve, move
around'. Now, obviously the coinage of a new word, such as
the Germanic one which comes out as English wheel, has to
depart from available linguistic material, but the
possibilities in any language for expressing an idea such as
a wheel are perhaps limited, but nevertheless so manifold
that in practice the contraints on the choice of an apt
expression seem to derive to a greater degree from general
constraints on conceptualization than from the specifics of a
given language.
In "(Micro-)Categorization, Semantic Change, and the
Sapir-Whorf Hypothesis" (pp. 91-103) Richard Rhodes also
discusses the Sapir-Whorf hypothesis in relation to lexical
change. He discusses three types of lexical changes. One of
them is the type where existing lexemes are applied to new
categories, as when the Algonquian language Miami applied
their old word for 'moose' to the 'deer' category when they
moved out of the range of the moose. Gy�ri would perhaps see
this as evidence that the Miami speakers are constrained by
their existing linguistic material and thus are more or less
forced to somehow conceptualize the deer as some kind of a
moose. Rhodes, however, draws upon these kinds of examples to
make the opposite point, arguing that linguistic determinism
cannot be literally true if the category to which a lexeme
applies can change independently of the lexeme. Both lines of
reasoning obviously cannot be right. It seems to be the case,
however, that both lack some nuances. Gy�ri is right that
lexical change often departs from existing vocabulary, but
his assumption that this automatically creates Whorfian
effects would be difficult to prove. For this reason Rhodes
is right in excluding cases of lexical recycling from the
purview of the Sapir-Whorf hypothesis, although his
description of the change of category as being independent of
the lexeme is not accurate, given that, as the case of
'moose' coming to mean 'deer' illustrates, the transfer of
category assignment is licensed by similarity in semantic
structure.
Another kind of lexical change drawn upon by Rhodes to
show that linguistic determinism cannot be true is that of
shifts where existing lexemes are applied to new categories,
but in this case categories that were already known to the
speakers of the language, e.g. as when the Ancient Greek word
for 'beech' comes to mean 'oak' as part of a chain shift
involving a series of tree names. Again it is a question
whether such shifts imply or do not imply Whorfian effects--
they are probably best viewed as being neutral to the
question.
The final kind of lexical change that Rhodes discusses
involves the co-evolution of words and categories. The
example given is of Ojibwe morphemes in the BREAK/TEAR domain
that seem to collectively develop a sensisitivity to the form
of the objects involved. Since Rhodes does not relate this
type of lexical change to the issue of linguistic relativity
it is not clear what this example is supposed to demonstrate.
Possibly the idea is that all lexical changes should exhibit
the kind of behavior as the Ojibwe morphemes in the
BREAK/TEAR domain if linguistic determinism were to be
acceptable.
On the whole Rhodes' paper is interesting for the
examples it brings into play, but it does not succeed in
demonstrating their direct relevance for the issue of
linguistic relativity.
In "Verbalized Events: A Dynamic Approach to Linguistic
Relativity and Determinism" (pp. 107-138) Dan Slobin reports
on work on the way that language influences the process of
formulating and interpreting verbal messages, specifically
departing from Leonard Talmy's typological observation that
some languages (so-called satellite-framed languages like the
Germanic ones) tend to encode manner descriptions in their
motion verbs while others (verb-framed languages like
Romance) encode path descriptions. Evidence from picture-
elicited narratives, fictional literature materials,
conversational data, translations involving contrasting
language types, and experiments where speakers of different
languages are asked to describe from memory a scene from a
book all confirmed the reality of the typological distinction
and the experimental data in particular lends credence to
Slobin's modified view of linguistic relativity. The paper
continues earlier work by Slobin where his version of
linguistic relativity issue is formulated in the words
"thinking for speaking." Slobin distances himself from the
more radical Whorfianism where languages are seen to
determine cultural patterns and world-view and sticks to the
more cautious assumption that the influence of language on
thought is a matter of degree of habitual attention.
Balhasar Bickel, who has been affiliated with the
Cognitive Anthropology Research Group at the Max-Planck
Institute in Nijmegen, explores Whorfian effects of spatial
deixis in the Tibeto-Burman language Belhare (Eastern Nepal).
In line with the results of some other researchers in the
Nijmegen group his paper "Grammar and Social Practice: On the
Role of 'Culture' in Linguistic Relativity" (pp. 161-191)
focuses on a language where the coordinates of deixis are
absolute rather than centered on the human body. The deictic
system, which pervades the grammar of Belhare, to the point
where even certain frequent interjections Aktionsart
modifiers, and case markers make deictic distinctions,
consistently distingish UP, DOWN, and ACROSS. This system is
contrasted with corporeal spatial expressions in Bickel's
native language Alemmanic, a language which is nevertheless
spoken in an environment that is topographically similar to
the Himalayan foothills, namely Switzerland. Bickel finds
that "[d]eixis relies on an awareness of structured space,
and at the same time, it creates such structure. Both these
aspects of deixis have clear repercussions for sociocultural
practices and experiences." (p. 176). In the case of Belhare,
this is demonstrated by examples from such areas as weaving,
"shamanistic" curing rituals, the organization of the
interior of a house, and ways of referring to people all of
which require an attention to the three absolute coordinates.
While Bickel does not go so far as to argue that the Belhare
deictic system determines its speaker's behavior thought or
behavior he does--convincingly--argue that it affects their
attention and plays a role in cultural practices.
The paper "Equivalence and Mismatch of Semantic
Features: Collocations in English, Spanish and Dutch" (pp.
29-51) by Jan Schroten does not directly deal with the
linguistic relativity hypothesis but rather with what is seen
by the author as a prerequisite for testing the hypothesis,
namely semantic analysis. Schroten presents some basic
features of Pustejovsky's compositional theory of semantics
and provides some examples of contrastive analyses, mostly
from the body-part domain, aimed at demonstrating the
usefulness of this approach.
In the paper "'S'Engager' vs. 'To show Restraint':
Linguistic and Cultural Relativity in Discourse Management"
(pp. 193-222) Bert Peeters explores the cross-cultural
pragmatics of French vs. English normative attitudes towards
linguistic interaction. Drawing upon a number of carefully
gathered examples and applying Anna Wierzbicka's cultural
scripts approach Peeters gives substance to common
observations about the differences between Anglo and French
ways of interacting. While the paper is interesting enough in
itself, the argument of the author that the kind of cultural
relativity that it discloses also implies linguistic
relativity is tenuous.
In her paper "Grammar and the Cult of the Virgin" (pp.
223-233) Elzbieta Tabakowska analyzes a number of Polish
expressions that conceptualize the Virgin Mary. Tabakowska
has chosen this domain in order to demonstrate how attitutes
of a given culture with respect to certain of its
characteristic elements are revealed and sustained by the
linguistic expressions pertaining to the cultural sphere in
question. Through lexical, morphological and syntactic
analysis she demonstrates that linguistic expressions mirror
the combination of the homely and the divine, the
transcendent and the mundane, which is seen as a particularly
Polish way of conceiving of the Virgin.
In conclusion, the book contains an interesting set of
papers. Some of them are only marginal to the problem of
linguistic relativity, but all relate explicitly to it in a
variety of ways and a few, in particular those of Imai,
Bickel, and Slobin, actually do succeed in providing some
evidence for linguistic relativity. Characteristic of all the
three papers just mentioned is that the authors are cautious
not to overstate the effect of language on thought and
behaviour: Imai emphasizes that "human cognition is neither
absolutely universal nor absolutely malleable by language"
(p. 158); Bickel stresses the mutual influence of language
and culture arguing that "sociocultural practices (...)
sustain the cognitive style and bias of awareneness that is
required by a particular grammar" (p. 185); and Slobin
remarks that "the followers of Humboldt, Sapir, and Whorf
have not succeeded in demonstrating a pervasive influence of
languages on world-voew or broad patterns of cultural
practices," but continues to add that "one cannot escape the
influences of language while in the process of formulating or
interpreting verbal messages." (p. 107). Approached in this
kind of cautious vein the Sapir-Whorf hypothesis gains
credence and will probably continue to be productive in the
future.
S�ren Wichmann is a Senior Researcher at the Danish Institute
for Advanced Studies in the Humanities and specializes in
synchronic and diachronic studies of Mesoamerican languages.
|
