时间：2022-07-07 03:24:47
Over the last 60 years， a wide range of theories about the relationship between language and thought have been developed and disputed. The theories that posit some shaping or determining effect of language on thought are typically grouped under the rubric of linguistic relativity theories， or more popularly， the Sapir-Whorf Hypothesis （SWH）.
Interestingly， the long-standing controversy over linguistic relativity has been only modestly impacted by two significant developments in our modern understandings of language and cognitionnamely， the now commonplace position that both language and cognition are fundamentally interactional and socially-situated practices that cannot be reduced to isolated， abstract knowledge structures.
The work in recent years on linguistic relativity by scholars such as John Lucy （1992， 1996） and Stephen Levinson （1996） represents a significant leap in theoretical sophistication and empirical rigor in comparison to earlier works. From the perspective I take here， however， there are important continuities between early and modern approaches. In particular， both early and modern approaches overwhelmingly analyze language in terms of the semantic properties and organization of a language’s lexicon and morph syntax. Consequently， theoretical and empirical attention has focused on the cognitive effects of the organization and semantic properties of color terms， kinship terms， and ethnobiological taxonomies， and on obligatory morphosyntactic categories such as noun and numeral classifiers， tense， and aspect. In these traditions， then， language is typically understood in terms of abstract structure， and not as discourse.
The dominant conception of cognition operative in both these traditions is similarly divorced from concerns with practice and interaction. In most work on linguistic relativity， cognition emerges as a process of categorization of objects of perception. This process is effectively theorized as being carried out by isolated individuals outside of real-life contexts of task-performance or social action. And although the linguistic and mental structures of interest are assumed to be replicated in the minds of all speakers of a given language， the existence of a community of speakers and of interactions between them has been of little theoretical or empirical importance.
The concept of distributed cognition arises from the observation that real-life cognition is rarely， if ever， a process bounded by the skull. Instead， real-life cognition involves interaction with other individuals， and with semiotic artifacts such as texts and maps. Research in distributed cognition therefore seeks to rethink the basic unit of cognition， expanding it beyond the brain to encompass the whole body， useful physical artifacts and technologies， and ultimately， groups of people. Among other things， then， distributed cognition concerns itself with the cognitive states and processes that can be ascribed to groups of interacting individuals （A. Clark 1997， Hutchins 1995）.
What is meant， though， by a group-level cognitive state or process？ An example can help here. In an early study of distributed cognition， Ed Hutchins examined the functioning of a navigation team aboard a naval vessel. The task of the navigation team is to keep track of the location of the vessel. It did this by taking measurements of the orientation of the vessel relative to visible landmarks. Using these data， a plotter triangulates the position of the vessel on a detailed navigational chart， a process known as “taking a fix”.
This process involves several individuals： two bearing-takers make and call out bearing measurements to another individual who records the information in a log. A fourth individual takes the information from the log and plots it on the navigational chart， and a supervisor manages the timing of the entire process. All of these individuals are involved in the flow of information that makes the fix possible， and the cognitive processing for the task is distributed among them. Working together， the navigational team is able to answer the basic question： “where are we？” Taking a fix in this way， then， is an example of a distributed cognitive process， a cognitive process that is fundamentally dependent on the interaction and contributions of multiple individuals.
Returning now to discourse， it will be readily appreciated that in the collaborative work of the navigation team， communicative interaction is crucial. Members of the group share information with one another， check their understanding and seek clarifications， and dispute and agree on interpretations by using speech and writing. Language enables members of the group to share aspects of their own cognitive processes with others， as well as to gain access to those of others. The result of this language-mediated interaction is not merely an aggregate of isolated individual cognitive processes but the emergence of a network of cognitive activity that supports group-level cognitive processes.
Any investigation of the relation between language and thought must also cope with “differences in patterns of use” in natural languages. The question is whether patterns of use have an impact on thought either directly or by amplifying or channeling any effects due to linguistic structure. We can call this the hypothesis of discursive relativity， a relativity stemming from diversity in the functional （or goal-oriented） configuration of language means in the course of （inter）action. （Lucy 1996， 52） Perhaps the first scholar to suggest a discourse approach to the linguistic relativity problem was Dell Hymes. In a series of papers in the 1960s （Hymes 1961， 1966）， Hymes proposed that while language structure may be important for questions of linguistic relativity， language use should be of even greater importance.
In the neo-Whorfian research of the 1990s， the notion of discursive relativity was further developed by several scholars， including Dan Slobin （1996）， John Haviland （1996）， John Gumperz （1996）， and Herbert Clark （1996）. Of this collection， Herbert Clark’s work on the achievement of joint attention in interaction， a process he calls coordination， stands out in its emphasis on a unit of analysisthe conversing pairthat at least implicitly challenges the assumptions of individuated cognition. However， with the exception of Clark’s work， and possibly that of the discourse- centered approach to culture， work on discursive relativity has remained committed to an individuated model of cognition. It is only with Ed Hutchins’ development of the concept of distributed cognition that the question of linguistic relativity with respect to units of cognition other than individuals receives significant attention.
When cognitive activities are distributed across social space， the language or languages used by task performers to communicate are almost certain to serve as structuring resources， and the structure of language will affect the cognitive properties of the group even if they do not affect the cognitive properties of individuals in the group. （Hutchins， 1995， p232）
In conclusion， then， I see the contribution of this paper as two-fold： first， it introduces an approach to cognition into the extant tradition of research on discursive relativityan approach which I believe has the capacity to significantly widen our understanding of how language and thought interact. And second， it introduces into the research on distributed cognition a discussion of how specific patterns of communication and specific discourse practices produce different distributed cognitive processes in interacting groups.
References：
[1]Clark， Andy. Being there： putting brain， body， and world together again. Cambridge： MIT Press， 1997.
[2]Clark， Herbert. Communities， commonalities， and communication. Rethinking linguistic relativity. Cambridge University Press， 1996.
[3]Gumperz， John and Stephen Levinson. Rethinking linguistic relativity. Cambridge University Press， 1996.
[4]Haviland， John. Projections， transpositions， and relativity. Rethinking linguistic relativity. Cambridge University Press， 1996.
（责编 张亚欣）