b. Goingbeyond individual items, languages havedifferentpatterns of lexicalization. For instance,Talmy (1980)
observes that Romance languages typically conflate path and motion together in a verb, while English
typically conflates manner and motion (though it has its share of path+motion verbs too). These differences
in pattern are correlated with differences in grammatical patterns as well.
c. Languages haveinflectional categories that reflectdifferentpartitioning of semanticspace—differentgender/
classifier systems, different tense/aspect systems, different politeness systems. Linguistic semantics must
enable us to characterize these differences.The answer to these arguments is the same as in the previous subsection. All of these arguments concern the way
elements of linguistic form map intocomplexesof meaning, not with the contents of meaning itself. Thus theycan be
characterized as language-specific differences in the interface rules—either (a) in the mapping associated with
particular cultural vocabulary, (b) in the general pattern of mappings encoded by classes of lexical items (which, recall,
are interface rules), or (c) in the phrasal interface rules associated with grammatical and morphological features. This,
then,is specificallylinguisticsemantics, notbecauseitinvokes a differentsort ofcognitivestructure,but ratherbecause
it involves how the vocabulary and grammar of different languages map onto thesamelevel of conceptual structure,
thereby creating different natural groupings of meanings for users of different languages.
Position (12d) is not far fro ma well-known position e merging fro mlinguistics and anthropology, often called the
Sapir–Whorfhypothesis (Carroll 1956). This stresses thedependence ofthoughtonlanguage, claiming thatdifferences
among languages strongly affect the thought processes of their speakers. Again there is a certain degree of plausibility
tothisclaim, particularlyin therealm of vocabulary. Infact,itis unnecessary tolook toother languages: we can simply
look to technical sub-vocabularies in our own language (say, chemical, medical, cultural, or religious terms) to see how
much greater precision is afforded in discourse and thought by virtue of having a morefinely divided vocabulary.^144
Whorf's more radical claim was that grammatical structure fundamentally affects thought. He claimed, for instance,
that theHopi language contains no elements that refer to time, and therefore that monolingual Hopi speakers have no
concept of time; both aspects of this claim have been refuted by Malotki
292 SEMANTIC AND CONCEPTUAL FOUNDATIONS
(^144) Incidentally, the“urban legend”that Eskimo languages have dozens of words for snow can be traced to a far less extreme claim by Whorf; the actual range is not that
different fro mEnglishsleet, slush, blizzard, powder, etc. (Pullu m1981).