Language Diversity and Social Action: A Third Locus of Linguistic Relativity
Author(s): Jack Sidnell and N. J. Enfield
Vol. 53, No. 3 (June 2012), pp. 302-333
Published by: The University of Chicago Press on behalf of Wenner-Gren Foundation for Anthropological
Stable URL: http://www.jstor.org/stable/10.1086/665697 .
Accessed: 15/05/2012 10:23
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact email@example.com.
The University of Chicago Press and Wenner-Gren Foundation for Anthropological Research are collaborating
with JSTOR to digitize, preserve and extend access to Current Anthropology.
302 Current Anthropology Volume 53, Number 3, June 2012
䉷2012 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved. 0011-3204/2012/5303-0003$10.00. DOI: 10.1086/665697
Language Diversity and Social Action
A Third Locus of Linguistic Relativity
by Jack Sidnell and N. J. Enﬁeld
The classic version of the linguistic relativity principle, formulated by Boas and developed especially in the work
of Whorf, suggests that the particular lexicogrammatical patterns of a given language can inﬂuence the thought of
its speakers. A second version of the argument emerged in the 1970s and shifted the focus to the indexical aspect
of language: any given language includes a particular set of indexical signs, and these essentially shape the contexts
produced in speaking that language. In this article, we propose a third locus of linguistic relativity. Our argument
is based on recent work in conversation analysis that has shown how the resources of a given language provide the
tools for accomplishing basic actions in interaction. To develop our argument, we consider the way in which the
resources of three different languages (Caribbean English Creole, Finnish, and Lao) are deployed by speakers to
agree with a prior assessment while at the same time claiming greater epistemic authority over the matter assessed.
Our case study indicates that the language-speciﬁc tools used to accomplish this action (the lexicogrammatical
resources) introduce collateral effects and in this way give the action a local spin or inﬂection.
Does speaking one language rather than another have con-
sequences for thought, and for social life more generally?
Anthropologists, linguists, philosophers, and psychologists
have all sought to answer this question. The enormous lit-
erature, spanning at least 200 years, encompasses everything
from formal logic and analytic philosophy to naturalistic ob-
servation, psychological experiment, and work bordering on
literature and ﬁction. Everyone from Edward Sapir to George
Orwell, from Franz Boas to Toni Morrison has had a say on
how the language one speaks does or does not affect one’s
understanding of, and place within, the surrounding world.
The question of linguistic relativity has been central to the
anthropology of language, although the methods for answer-
ing the question and the kinds of answers given have un-
dergone signiﬁcant transformation. Below, we begin by brieﬂy
reviewing two broad conceptions of linguistic relativity—ﬁrst,
the notion that different languages can have different effects
on thought (e.g., conceptual representations and inference),
and second, the notion that different languages can have dif-
ferent effects on sociocultural context (e.g., social relations
among interlocutors)—noting some recalcitrant methodo-
logical and analytic issues, before turning to the main aim of
Jack Sidnell is Associate Professor in the Departments of
Anthropology and Linguistics at the University of Toronto (19 Russell
Street, Toronto, Ontario M5S 2S2, Canada [jack.sidnell@utoronto
.ca)]. N. J. Enﬁeld is Professor of Ethnolinguistics at Radboud
University Nijmegen (PB 310, 6500 AH, Nijmegen, The Netherlands
[nick.enﬁeld@mpi.nl]). This paper was submitted 2 X 10 and
accepted 28 II 11.
the paper. Our goal is to formulate a third version of the
linguistic relativity argument grounded in situated social in-
teraction, a realm in which our behavior is not primarily about
exchanging information but about getting things done in and
through our social relations. In this third locus of linguistic
relativity, different languages can have different effects on the
kinds of social actions that can be achieved through social
Synopsis of the Argument to Be Presented
Within the existing anthropological literature, we can discern
at least two distinct versions of the relativity argument. The
ﬁrst—associated with the founders of linguistic anthropol-
ogy—links language-speciﬁc patterns of grammar to thought
(in the general sense of mental representations of states of
affairs and the inferences that arise from these representa-
tions) and habitual behavior (Boas 1911; Sapir 1921, 1949
, 1964 ; Whorf 1956 , 1956 , 1956
). This version has received signiﬁcant attention from
linguists and psychologists and, over the past 25 years or so,
has been pursued in a wide range of psychologically informed
studies that use experimental methods to test the cognitive
consequences of language diversity (Boroditsky 2001; Gentner
and Goldin-Meadow 2003; Lucy 1992a, 1992b, 1997; Majid,
Boster, and Bowerman 2008; Pederson et al. 1998; Winawer
et al. 2007, inter alia). Studies in this vein have become in-
creasingly sophisticated in terms of both method and their
grounding in linguistic typology. At the same time, their per-
Sidnell and Enﬁeld Language Diversity and Social Action 303
ceived relevance to social and cultural anthropology has less-
ened because of the increasing focus on individual psychology
rather than culture or linguistic practice.
A second version of the relativity argument was hinted at
by Hymes (1966), made explicit by Silverstein (1976, 1979),
and subsequently elaborated by a number of linguistic an-
thropologists (see, e.g., Agha 1994, 2007; Briggs 1986; Er-
rington 1985, 1988; Hanks 1990; Ochs 1988, 1990, 1992, 1996;
Rumsey 1990). In contrast to the earlier emphasis on refer-
ential and predicational aspects of language, Silverstein fo-
cused on the indexical relations between speech and its con-
text of occurrence. Indexicality is a sign-function in which a
signiﬁed is linked to a signiﬁer by a relationship of contiguity
(classic examples include smoke being taken to signify ﬁre
and a knock being taken to indicate someone at the door;
see Kockelman 2005; Parmentier 1994a, 1994b; Peirce 1955).
The basic argument is that by the very act of speaking, speak-
ers both indexically presuppose and create, moment-by-
moment, a context. Thus, in saying “Do you know the way
to Conn Hall?” the speaker indexically establishes (among
other things) an addressee (a “you”).
Now consider the dif-
ference between English and French. In French, a speaker
must choose between address with tu or vous (and of course,
if you say tu you can be heard as not saying vous, and vice
versa), whereas in English there is only one second-person
pronoun (Brown and Gilman 1960). These different forms
convey something about the relationship between speaker and
addressee and about the context in which the talk takes place.
Within some speciﬁc set of contextual presuppositions, every
act of address in French thus takes aspects of the relationship
between speaker and hearer and makes them explicit, yet these
same aspects need not be articulated in English at all.
difference between the two languages appears to have con-
sequences for the contexts that their speakers establish
through speaking one language or the other. It is a simple
example because the range of alternatives (tu vs. vous)isso
narrow. Things become considerably more complex when we
consider, as we do below, person reference and address in
Vietnamese, where the range of alternatives is extensive and
where a number of “perspective-taking” strategies are also
used. The argument may also be applied to other domains
and not just the social deixis of tu-vous-type alternations or
Vietnamese person-referring forms. Even more basic, perhaps,
are the indexical signs (primary among them deictics such as
“here” and “now”) by which participants convey and thus
constitute or construe the here-and-now of any actual social
1. Let us clarify what this means. When I ask you a question, and
because I ask it, you become an addressee. Obviously, my question has
not created you, but it has created a status (in the sense of a set of
entitlements and responsibilities) that you now ﬁll (e.g., you are now
accountable for not answering the question).
2. Explicitness is important because it relates to whether speakers po-
tentially “go on record” and may thus make them accountable for what
they have said or done. This can be seen in relation to T-V (tu-vous)
address forms in Jacquemet (1994).
encounter. If the ﬁrst version of the relativity argument em-
phasizes the consequences of language diversity for the world
perceived, the second focuses on the world indexed (and thus
produced) in different ways through different languages, in
and through the very act of speaking.
To these now well-established versions of the relativity ar-
gument we want to add a third. The ﬁrst version began with
language conceptualized as a system for thought and the sec-
ond with speaking as meaningful social behavior.
version begins with practices of social interaction and the
particular forms of social action that they provide for.
thesis is that different grammatical and lexical patterns of
different languages can provide different opportunities for
social action. Since Wittgenstein (1953) and Austin (1962),
it has been recognized that by speaking we are not simply,
solely, or primarily engaged in describing the world, depicting
it, or indexing it in some way. Rather, by speaking we are
acting in it. When you say “That’s a really nice jacket,” you
have not only described someone’s clothing; you have given
them a compliment (see Pomerantz 1978). When you say
“The trafﬁc was terrible today,” you have not only described
your commute; you have complained about it (Drew and
Walker 2009). When you say “Could you give me a lift?” you
have not only asked a question; you have made a request
(Curl and Drew 2008).
An empirical grounding for this action-based approach can
be found in research in conversation analysis. This research
tradition focuses on the practices of speaking (and behavior
more generally, e.g., gesturing) deployed to accomplish par-
ticular social actions. Such actions are organized into se-
quences in such a way as to establish “an architecture of
intersubjectivity” or “a continuously updated context of in-
tersubjective understanding” (Heritage 1984:254–260; see
also, inter alia, Atkinson and Heritage 1984; Goodwin and
Heritage 1990; Levinson 1983; Sacks 1995; Schegloff 1968,
2007; Sidnell 2010). When we examine conversation, we see
that participants themselves can use subsequent turns at talk
as evidence for whether and how they have just been under-
stood by the other. This participants’ method can also be
exploited by researchers as a methodological and analytical
lever. Thus, to determine the particular action that some bit
of talk achieves, we can look to see how it was taken up (or
not) in subsequent talk (see Sacks, Schegloff, and Jefferson
3. Silverstein (1976:11) begins, “This chapter will try to develop con-
sequences of the statement that speech is meaningful social behavior.”
4. Silverstein (1987) makes this same distinction between indexicality
and purposive action. Our “third locus” relates to Silverstein’s “func-
,” which concerns “the purposive, goal-oriented use of speech (or
equivalents) by intentional individuals in speciﬁc situations of discourse,
each such usage constituting a ‘speech act’ or ‘speech event.’” (Silverstein
1987:23). His “function
,” in contrast, “consists of multiple relationships
of existential implication among isolable elements/aspects of a com-
municative situation. In particular, we can see linguistic elements as the
principal system of indexicals, the elucidation of which is a third kind
of functional explanation” (Silverstein 1987:31). This is the domain of
the “second locus” of linguistic relativity discussed above.
304 Current Anthropology Volume 53, Number 3, June 2012
1974:728–729 on the so-called next-turn proof procedure).
This is crucial because it solves one of the more difﬁcult
methodological problems of all relativity arguments (see Lucy
1992a, 1992b): how to show that a grammatical or lexical
peculiarity has nonlinguistic (i.e., cognitive, cultural, action-
relevant) consequences. The ﬁrst version of the relativity ar-
gument has come to rely on various forms of experiment
using measures such as memory and inference to demonstrate
such consequences. The second version typically relies on
native-speaker testimony and ethnographic description. A fea-
ture of the third version is that the consequences,while clearly
nonlinguistic, are nevertheless internal to the data.
We illustrate this third approach to linguistic relativity by
investigating the different ways in which a speciﬁc type of
social action is carried out, using as a vehicle the lexicogram-
matical resources of three different languages: Caribbean En-
glish Creole (Sidnell 2009c), Finnish (Hakulinen and Sorjonen
2009; Sorjonen 1996; Sorjonen and Hakulinen 2009), and Lao
(Enﬁeld 2007a). The idea is to use the controlled comparison
of a single general type of action (what we call an “episte-
mically authoritative second-position assessment,” deﬁned
below as the action of agreeing with what someone has just
said while simultaneously signaling that one has greater au-
thority to have said it; see Heritage and Raymond 2005).
Each case study reveals the speciﬁc grammatical resources
employed in that language as well as the associated interac-
tional consequences. A lexicogrammatical structure will be an
appropriate tool or vehicle for carrying out a certain social
action so long as its semiotic affordances make it well suited
to effecting that action, for example, to the extent that people
will recognize the function it is being used for. But because
any such lexicogrammatical structure will have other struc-
tural properties as well (including other meanings), other
semiotic affordances are unavoidably introduced. The struc-
ture is selected because it has a certain functional feature, but
other properties of that structure will be ushered in, and these
collaterally selected “features,” in turn, either will be features
for other functions or may turn out to be not features but
bugs. Either way, the selection of a linguistic structure based
on one feature will inevitably introduce other features that
give rise to what we refer to as collateral effects, that is, side
effects of the selection of a speciﬁc means for some ends (see
Our three example cases are useful for a number of reasons.
First, a good deal is known about the phenomenon central
to these cases: “responses to assessments” (see esp. Goodwin
and Goodwin 1987; Heritage and Raymond 2005; Pomerantz
5. To be clear: the action here is “agreement.” The “epistemically au-
thoritative” part of what we describe here is laminated onto agreement.
So it would, of course, be quite impossible to merely claim epistemic
authority without embedding this in some action (e.g., agreement). It is
perhaps reasonable to describe, as Heritage and Raymond (2005) do,
epistemic authority as being “indexed” here. This would suggest that our
case study, while focusing on action, also introduces some features of
the second version of relativity we discuss.
1984). Second, the languages differ signiﬁcantly from each
other in structure. Both Lao and Caribbean English Creole
are highly analytic languages largely devoid of any inﬂectional
morphology. Finnish, by way of contrast, has an extensive set
of inﬂectional morphemes (sufﬁxes) that attach to both nouns
and verbs: nouns are inﬂected for case (nominative, accusa-
tive, partitive, genitive, locative, etc.); verbs are inﬂected ac-
cording to the person and number of the subject; and other
verbal sufﬁxes convey distinctions of tense, aspect, mood, and
so on. While Caribbean English Creole uses word order to
convey grammatical relations or semantic roles such as agent
and patient, in Finnish, word order is relatively free and these
relations are expressed by inﬂections on nominal arguments.
Lao is like the Caribbean Creole in being highly analytic, but
it differs—crucially, for our purposes—in that it has an elab-
orate system of ﬁnal particles.
Boas to Whorf: Linguistic Relativity in
A classic version of the relativity hypothesis takes language-
speciﬁc grammatical patterns and relates them to patterns of
thought or, in more recent terms, cognition (of the many
recent studies, see, e.g., Everett 2005; Levinson 2003a, 2003b;
see also Gumperz and Levinson 1996 for a useful state-of-
the-art volume as of the mid-1990s). Although the ideas can
be traced to German Romantics such as Herder (2002 )
and Humboldt (1999 ), this version found full ex-
pression for the ﬁrst time in the writings of Franz Boas (see
Duranti 1997, 1999, 2003, 2009; Leavitt 2006; Lucy 1992b;
Sahlins 1976; Stocking 1996). The introduction to The Hand-
book of American Indian Languages (Boas 1911) contains his
most famous statement of the relativity argument.
As part of a broader program of anthropological research,
Boas developed an approach to language structure—in par-
ticular, grammatical categories—that, like other forms of
structural analysis, drew attention to the internal relations
between components of lexical and grammatical subsystems
6. Boas developed an earlier version of the relativity argument in
relation to the perception of speech sounds in an 1889 article titled “On
Alternating Sounds” (Boas 1889). Several Europeans had reported that
in certain “Indian” languages, speakers did not make clear and regular
distinctions between sounds. On one occasion a sound would be realized
as “nd” and on another as “tl.” Boas explained that the phenomenon of
alternating sounds lay not in the observed phenomenon but rather in
the observer’s own preconceptions. The argument is summarized in the
1911 introduction to the Handbook, where Boas (1911:17) writes,
Thus, the lower Chinook has a sound which is readily perceived as a b,
m,orw. As a matter of fact, it is a bsound, produced by a very weak
closure of the lips and with open nose, the breath passing weakly both
through the mouth and through the nose, and accompanied by a faint
intonation of the vocal chords....observers belonging to different
nationalities readily perceive the sounds in accordance with the system
of sounds with which they are familiar....Thealternation of the sounds
is clearly an effect of perception through the medium of a foreignsystem
of phonetics, not that of greater variability of pronunciation than the
one that is characteristic of our own sounds.
Sidnell and Enﬁeld Language Diversity and Social Action 305
so as to reveal their partially arbitrary, could-have-been-
otherwise character (see Jakobson 1971  on Boas’s ap-
proach to grammatical meaning). Both Kwak’wala and Inuk-
titut, the languages Boas studied most intensively, exhibit
complex patterns of phonology, grammar, and word forma-
tion associated with the “polysynthetic” language type (Com-
rie 1981b; Sapir 1921). They also include grammatical cate-
gories (such as obviative, hearsay evidentials, and deictics that
convey whether the referent of a noun phrase is visible to a
speaker) that were largely unfamiliar to the philological tra-
dition of linguistic analysis that had dominated until the turn
of the twentieth century.
Grammatical categories encode distinctions such as tense,
aspect, person, number, and deﬁniteness. Languages differ in
terms of both which grammatical categories they include and
how those included are organized by relations of opposition
and difference. Boas argued that these differences were con-
sequential not so much for what they allowed a speaker to
say as for what they required a speaker to say. For instance,
in order to produce a grammatically well-formed sentence in
Kwak’wala, a speaker is required to indicate how they know
what they are asserting (whether they saw it, heard it, heard
about it, dreamt it, etc.; see Aikhenvald 2004). This leads
directly into the issue of relativity, since in forming a sentence
in Kwak’wala, a speaker is literally forced by the requirements
of grammatical well-formedness to attend to certain aspects
of reality not demanded of a speaker of English. These pro-
posals have been pursued in recent work on “thinking for
speaking,” where it has been suggested that “language directs
us to attend—while speaking—to the dimensions of experi-
ence that are enshrined in grammatical categories” (Slobin
1996:71). Thus, for example, if our language requires us to
encode singular versus plural, we are more likely to pay at-
tention to “plurality” in scenes that we will later need to
Boas contrasts word-formation processes in English with
those in other languages, alluding to the polysynthetic patterns
of Kwak’wala and Inuktitut:
The groups of ideas expressed by speciﬁc phonetic groups
show very material differences in different languages, and
do not conform by any means to the same principles of
classiﬁcation. To take again the example of English, we ﬁnd
that the idea of water is expressed in a great variety of
forms: one term serves to express water as a liquid; another
one, water in the form of a large expanse (lake); others,
water as running in a large body or in a small body (river
and brook); still other terms express water in the form of
rain,dew,wave, and foam. It is perfectly conceivable that
this variety of ideas, each of which is expressed by a single
independent term in English, might be expressed in other
languages by derivations from the same term. (Boas 1911:
In a polysynthetic language such as Kwak’wala, it is possible
for a single lexical root (e.g., that meaning “water”) to com-
bine with a great number of inﬂectional morphemes to pro-
duce a range of semantically diverse words. Boas seems to
have been suggesting that, by virtue of their common use of
a single identiﬁable root, the ideas expressed are grouped
together in a special way. Common roots invite analogical
thinking about objects in the world that from the perspective
of the lexicon of another language may appear quite disparate
(see Lucy 1992b).
Another passage focuses on grammatical categories. Boas
notes that languages differ, ﬁrst, in terms of what grammatical
categories they include and, second, in how these grammatical
categories are conﬁgured:
When we consider for a moment what this implies, it will
be recognized that in each language only a part of the com-
plete concept that we have in mind is expressed, and that
each language has a peculiar tendency to select this or that
aspect of the mental image which is conveyed by the ex-
pression of the thought. To use again the example which I
mentioned before, The man is sick. We express by this sen-
tence, in English, the idea, a deﬁnite single man at present
sick. In Kwakiutl this sentence would have to be rendered
by an expression which would mean, in the vaguest possible
form that could be given to it, deﬁnite man near him invisible
sick near him invisible. Visibility and nearness to the ﬁrst or
second person might, of course, have been selected in our
example in place of invisibility and nearness to the third
person. (Boas 1911:43; see also the related discussion in Boas
Boas is highlighting differences in the conﬁguration and
distribution of grammatical categories. Whereas in the English
example tense and deﬁniteness are obligatory, the sentence
from Kwak’wala requires the speaker to include information
pertaining to the location of the person talked about and his
visibility or nonvisibility to the participants in the speech
The examples could be expanded more or less indeﬁnitely,
drawing on decades of empirical work on the diversity of
grammatical systems around the world, work that has inten-
siﬁed in recent years (see Dixon 2009; Haspelmath et al. 2005;
Shopen 1985, 2007). In The Languages of Native North Amer-
ica, for instance, Mithun (2001) catalogues a range of gram-
matical categories, including those pertaining to “obviation”
(the so-called fourth person of, e.g., Ojibwa), “inverse num-
ber,” “distributives,” “associatives,” “locational and direc-
tional deixis,” “evidentiality,” and so on. The sheer degree
and richness of the diversity here is overwhelming, and if one
wants further evidence, the many available surveys of the
world’s languages provide a wealth of it (see, inter alia, Adelaar
and Muysken 2007; Comrie 1981a; Dixon 2002; Dixon and
Aikhenvald 1999; Foley 1986; Krishnamurti 2003; Posner
1996; Shibatani 1990; Sua´rez 1983).
Cognitive consequences of linguistic diversity. While the facts
of linguistic diversity are well established, its cognitive rele-
306 Current Anthropology Volume 53, Number 3, June 2012
vance and consequences remain open and hotly disputed.
was Benjamin Lee Whorf who developed this aspect of Boas’s
argument most explicitly. Whorf’s arguments and studies have
spawned a massive and somewhat unruly secondary literature
(for reviews, see, inter alia, Hill and Mannheim 1992; Hunt
and Agnoli 1991; Kay and Kempton 1984; Koerner 1992; B.
Lee 1985; P. Lee 1991, 1996, 2000; Lucy 1985, 1992b, 1996,
1997; Lucy and Shweder 1979). Whether or not one is con-
vinced by Whorf, he must be credited for pushing the rela-
tivity argument forward by insisting on evidence of the cog-
nitive or behavioral consequences of grammaticaldifferences.
Where Boas (1911:43) was merely suggestive—“each language
has a peculiar tendency to select this or that aspect of the
mental image which is conveyed by the expression of the
thought”—Whorf went farther, seeking to ground his argu-
ments about relativity in an ethnography of Hopi daily life.
While few scholars are today convinced by Whorf’s efforts in
this direction, many have followed his lead in attempting to
provide evidence for the consequences of grammatical dif-
ferences on the nonlinguistic behavior of speakers. This has
naturally encouraged experimental approaches that adopt a
broadly psychological perspective (e.g., Berlin and Kay 1969;
Brown and Levinson 1993; Gentner 1982; Gentner, Imai, and
Boroditsky 2002; Imai and Gentner 1993; Levinson 1992,
1996a, 1996b, 1997a,1997b, 2000, 2001, 2003a,2003b).
Some persistent issues of method seem to inevitably arise
in relation to this line of research. Consider one of the most
compelling research projects in this vein—associated with Ste-
phen C. Levinson and colleagues at the Max Planck Institute
for Psycholinguistics—focusing on the cognitive conse-
quences of differences between languages in the linguistic en-
coding of spatial location (Levinson 2003a; Majid et al. 2004;
Pederson et al. 1998). One key distinction is between lan-
guages that differ in terms of whether their speakers prefer
to use absolute versus relative spatial-reckoning systems (see
also Haviland 1993, 1996, 1998, 2000; Levinson 1996c). So a
speaker of English might say “Shift to the left,” while a speaker
of Guugu Yimidhirr would say the equivalent of “Shift to the
north.” Consequences of this linguistic difference for thinking
were tested in an experimental comparison between speakers
of Dutch (which, like English, uses a relative spatialreckoning
system) and speakers of Guugu Yimidhirr (which uses an
absolute system). In one experiment, a participant sat at a
table on which three toy animals were laid out in a row, all
7. For a different debate raised by the empirical facts of linguistic
diversity—that is, what kind of cognitive mechanism underlies it and
makes it possible—see Evans and Levinson (2009), Levinson and Evans
(2010), and commentary on these.
8. Two of Whorf’s other contributions deserve mention here. First,
Whorf conceptualized the matter not in terms of isolated grammatical
items/subsystems, as Boas had tended to, but in terms of broad collections
of quite disparate features that together constituted “fashions of speak-
ing.” Second, Whorf introduced the notion of cryptotypes orcryptotypic
categories—that is, categories that may not have any single, formal re-
alization in all contexts (see Lee 1996, Lucy 1992b, and Silverstein 1979
facing in the same direction. The participant was asked to
remember the position of these objects and was then rotated
180⬚so that he was facing in the opposite direction. He was
handed the same three toy animals and was instructed to lay
them out exactly as he had found them on the other table.
A majority of the Guugu Yimidhirr speakers adopted an ab-
solute strategy and replaced the animals so that the one that
was in the northernmost position on the ﬁrst table was again
in the northernmost position on the second table, and so on.
Furthermore, these Guugu Yimidhirr participants placed the
animals so that they were facing in the same (cardinal) di-
rection as before. They did not relate the positions of the
animals to their own location or spatial orientation. A ma-
jority of the Dutch speakers, by way of contrast, arranged the
animals so that they were in the same position relative their
own location, now adjusted; that is, the one that was to the
speaker’s left in the ﬁrst array was also placed to the speaker’s
left in the second array (see ﬁg. 1).
This study, and many like it, offers strong evidence that
language-speciﬁc patterns of grammar and lexicalization may
signiﬁcantly restructure cognition. Such work is part of a
broader complement of approaches to the linguistic relativity
problem. The arguments initially formulated by Boas, Sapir,
and Whorf were framed in terms of the signiﬁcance of lan-
guage diversity for thought, perception, and habitualbehavior.
In an effort to operationalize these ideas and ground them
empirically, researchers have adopted an experimental meth-
odology and in so doing have signiﬁcantly limited the scope
Experimental methods work with proxies of var-
ious kinds. For instance, in the example discussed above,
memory is a proxy for “thought” and possibly also for “ha-
bitual behavior.” Still, the experimental methods of a (broadly
speaking) cross-cultural psychology have led in many cases
to a private and individual view of mind, whereas the argu-
ments of Boas, Sapir, and Whorf were associated with a social
view of mind anchored in the collective representations of a
The view of mind implicit in the writings of Boas
and Whorf was an essentially cultural and anthropological
one that did not privilege the individual. “Thought” in the
phrase “language and thought” was not simply the sum of
individual memory, reasoning, and inference. Grammatical
differences and variation in patterns of lexicalization—
Whorf’s “fashions of speaking”—were linked to distinctive
cultural patterns, that is, to the very characteristics that made
9. See Lee’s (1996) critique of the neo-Whorﬁan movement as rep-
resented, in particular, by Lucy (1992a, 1992b). With special attention
to the work of Whorf, Lee contends that linguistic relativity began as a
“principle” to be argued for but is now (wrongly, she argues) beingtreated
as a “hypothesis” to be tested (see also Lee 2000).
10. But see Boster (1985, 1986); Boster and Johnson (1989); Boster,
Johnson, and Weller (1987); Romney (1999); Romney, Batchelder, and
Weller (1987); and Romney, Weller, and Batchelder (1986) for extensive
attention to the problem of consensus in cognitive anthropology.
Sidnell and Enﬁeld Language Diversity and Social Action 307
Figure 1. Recall task: animals in a row; ﬁgure 4.11 from Levinson (2003b:156).
the Kwak’wala the Kwak’wala and the Hopi the Hopi (see
Behavioral experiments produce measures of performance
on a particular task but leave us with the question of how
such performance measures relate to the ordinary activities
and thinking of the people in question. The methods isolate
the phenomenon of interest from the contexts and activities
within which it normally operates. They assume that to ab-
stract thought from those contexts does not fundamentally
alter its character. However, research on the distributed nature
of cognition calls this into question (see research described
in Clark 2007; Goodwin 2000, 2003, 2006; Hutchins 1995,
2006; Lave 1988; Lave and Wenger 1991, inter alia, showing
the ways in which thinking happens through artifacts/tools
and conﬁgurations of persons).
Silverstein’s Reformulation: Relativity
Silverstein’s reformulation of the linguistic relativity issue
emerged in the 1970s and was consolidated in a wide range
of studies produced both by him and by his students and
11. Levinson (2003b) goes in this direction, especially in chapter 4,
titled “Absolute Minds: Glimpses into Two Cultures.” But because the
view we are offered is so focused on the particular issue of spatial coding,
we do not have a broader sense of everyday life in these communities.
This is not the goal of Levinson’s book; as Hanks (2006) points out in
his review, the lack of sustained ethnographic attention is reﬂective of
the fact the book is not really directed to anthropology.
colleagues (e.g., Silverstein 1976, 1979, 1981, 1985, 1987, 2003,
2004). Our thumbnail summary in this section exempliﬁes
the central points by drawing on two quite different studies,
of Vietnamese person reference and of Mayan deixis.
Silverstein (1976) argues that anthropologists have mistak-
enly taken language in its referential-predicational (i.e.,prop-
ositional) function as the model for culture in general, when
in fact the referential-predicational function of language is
what makes it unique with respect to all other media of cul-
tural communication and signiﬁcation (see also Silverstein
1987). According to Silverstein, the key to understanding the
multifaceted relations between language use and social life lies
in an understanding of the indexical mode of signiﬁcation.
This is the mode by which a sign can stand for something
because it is connected somehow (e.g., in time or place) with
that thing. There are two basic modalities. On the one hand,
language use may presuppose certain aspects of context. This
is true in the relatively trivial sense that any speech signal
presupposes some source (e.g., a speaker) and in the nontrivial
sense that speciﬁc aspects of context must exist cognitively
and/or physically if speech is going to be properly interpret-
able. On the other hand, language use may indexically create
aspects of context (the rights and duties of an addressee, an
audience, a key, etc.). Given this dual or dialogic relationship
of contiguity, we can say that language is, in an important
“Creative” uses of language are those in which a referential
index serves to “make explicit and overt the parameters of
308 Current Anthropology Volume 53, Number 3, June 2012
Table 1. Simpliﬁed synopsis of Mayan deictics (adapted from Hanks 1990:18–19)
id base aʔoʔb’eʔiʔeʔØ Gloss
he´ʔe(l) he´ʔel aʔHere it is (Tact Pres)
he´ʔe(l) he´ʔel oʔThere it is (Vis Dir)
he´ʔe(l) he´ʔeb’eʔThere it is (Aud Dir)
teʔe(l) te´ʔel aʔRight there, here (Immed)
teʔe(l) te´ʔel oʔThere (Non-Immed)
way way eʔ(In) here (Incl)
to(l) tol oʔ(Out) there (Excl)
le(l) lel aʔThis one (Immed)
le(l) lel oʔThat one (Non-Immed)
le le tiʔThe one
le le The (def art)
structure of the ongoing event” (Silverstein 1976:34). Index-
ical pronouns, for instance, have a creative function in setting
the roles of speaker and the hearer versus addressee. For Sil-
verstein, the more creative indexes include those that signal
social distance, hierarchy, or, conversely, solidarity (e.g., tu-
vous-type distinctions in Indo-European languages; Brown
and Gilman 1960), which, he claims “by their very use, make
the social parameters of speaker and hearer explicit” (Silver-
stein 1976:34). Further, all indexes “range on a sliding scale
of creativity or performativity value from the extreme pre-
supposition displayed by deictics to the extreme creativity
displayed by subtle social indexes” (35). Silverstein proposes
a second axis of classiﬁcation distinguishing referential in-
dexes (such as ﬁrst- and second-person pronouns and de-
monstrative deictics) from nonreferential ones (such as
“brother-in-law” lexicons, social sex markers, and deference
indexes of speaker-hearer relations).
For present purposes, one central argument from Silver-
stein (1976) may be summarized as follows: different lan-
guages include quite different collections and conﬁgurations
of indexical signs that are activated in speaking. These in-
dexical signs help to constitute the context within which any
bit of speech signal can occur. Thus, in speech, different lan-
guages constitute differently conﬁgured contexts, and thus,
because the lived world centrally involves persons interacting
with one another and thus speaking to one another, ultimately
different sociocultural worlds arise. So, for instance, the rel-
evance of a kinship system that divides the world into clas-
siﬁcatory brothers-in-law on the one hand and everybody else
on the other is made manifest in every act of speaking a
12. As Silverstein points out, many forms combine referential and
nonreferential indexicality—so, for instance, the French tu combines a
referential function of address (“referring,” as it were, to the addressee)
while at the same time conveying, nonreferentially, greater “familiarity”
language like Guugu Yimidhirr (see Haviland 1979; cf. Enﬁeld
Hanks (1992:48) writes that a natural language deﬁnes in-
teractive context “by encoding pragmatic categories and forms
of interaction in the grammar itself” (see also Hanks 1993,
2005a). He develops this argument through a detailed study
of deixis in Yucatec Mayan. Deictic forms in Yucatec Mayan
are often composed of two morphemes, a base, which Hanks
calls an initial deictic (id), and a sufﬁxal or enclitic element
labeled a terminal deictic (td). Table 1 shows a selection of
these bimorphemic deictic forms. ids are displayed along the
vertical axis and tds along the horizontal one.
Deictic constructions occur in two main surface shapes:
continuous (examples , ) and discontinuous (, ;
(1) he´ʔel aʔ
Here it is (presenting).
(2) ¢’a´ah te´ʔel oʔ
Vb-imper dloc td
Put it there.
(3) he´ʔel a maa´skab’ aʔ
ostev Pro2 ntd
Here’s your machete (presenting).
(4) ¢’a´ah te´ʔich k’oo´b’en oʔ
Vb-imper dloc Prep ntd
Put it there in (the) kitchen.
ids include a grammatical-category feature (adverb, loca-
tive, etc.) and a “domain” feature, which speciﬁes the di-
mension of deictic frame signaled about. tds, on the other
hand, include no grammatical-category feature and cross-cut
the domain conveyed by the id, specifying “relative proximity
of referent to ground.” tds differentiate “modes of access”
Sidnell and Enﬁeld Language Diversity and Social Action 309
within a domain, for example, relative immediacy within
space or relative fullness of perception (tactual, visual, pe-
Hanks (1990:400–462) provides a detailed case study of the
use of Yucatec Mayan spatial deictics:
Inclusive way e
Here (around me)
Exclusive tol o
There (out, excluding me)
Sociocentric immediate te´?el a
Sociocentric nonimmediate te´
With one or two exceptions, inclusive way e
form a contrast pair turning on whether the region
referred to does or does not include the speaker. This means
that Yucatec Mayans do not show a guest where to sit by
saying the equivalent of “Sit here” (unless the speaker is oc-
cupying that very location at the time of speaking). Nor do
Yucatecans, anticipating a future location, suggest “Let’s go
here” (e.g., indicating a way through the brush; see Hanks
1990:409). Thus, although way e
occupies, in the Mayan
system, a position roughly analogous to that of English “here,”
Hanks shows that it differs signiﬁcantly in terms of possible
referential extension. And while way e
trasts with tol o
in discourse, Hanks shows that a number
of other semantic oppositions are also possible.
Differences between the Yucatec Mayan and English sys-
tems are even more obvious when we consider sociocentric
forms. Hanks provides the following example, in which he is
called to the table for lunch:
Come right here Will. Let’s eat. (BB.5.22)
Hanks (1990:425) writes, “He proceeded to offer me a seat
next to him at the table. . . . it was not his own seat he was
offering me, since if it were, his deictic reference would have
been phrased in egocentric terms.” He also tells us that ko´
, “Let’s go here,” is a “perfectly routine utterance
used to lead an addressee along a path.” Summarizing, Hanks
(1990:425) notes, “unlike the Inclusive ‘here,’ where a speaker
is, the Immediate ‘here’ is place to which (s)he can go.” The
is also the form used to segment the body,
for instance, in indicating where one is feeling a described
(6) Bey tu´uu´c ih te´
So then it happened right here like this.
The relevance of this to our argument should by now be clear.
These deictic forms grammaticalize the “space” within which
any interaction takes place. Thus, in using any one of these
forms, the speaker is construing and constructing the very
context within which that interaction is taking place and thus
is bringing to bear on the current situation the historically
constituted language system that is Yucatec Mayan. Differ-
ences in the deictic systems of Yucatec Mayan and, for in-
stance, English can thus be seen to result in differently struc-
tured contexts. Deixis, then, involves the intersection of the
historically constituted linguistic code (i.e., culture) and the
subjective experience of the individual. In a larger argument
spanning several works, Hanks (see especially Hanks 2005a,
2005b) develops the idea that one’s very sense of one’s body,
one’s place in the world, and one’s experience of the world
is shaped by the particular grammar of deixis one has cul-
Vietnamese Address and Self-Reference
Another example of the indexical formulation of linguistic
relativity is provided by research on Vietnamese person ref-
erence and address (see Luong 1984, 1987, 1988, 1990). In
Vietnamese reference and address, speakers routinely avoid
the use of pronouns in favor of kinship terms and names, so
that instead of saying “I saw you at the market,” one might
say the equivalent of “Niece saw Uncle at the market” (de-
pending, of course, on the relation between speaker and ad-
dressee). Drawing on Luong’s study, Agha (2007:356–357)
argues that such self-referential and vocative uses of kin terms
are “denotationally anomalous in that they employ third per-
son nouns in referring to speaker and addressee (for which
ﬁrst and second person pronouns are available).”
forms are then used for non-kin, as they routinely are, a
further trope comes into play—one that suggests hierarchical
kin relations as a model for all social relations.
An additional complication is introduced by the possibility
of various shifts of footing (Goffman 1981), in which a
speaker adopts the perspective of another person. In some
cases, such perspective-taking strategies are normatively re-
quired. Luong (1990:57) writes, for instance, that “in his in-
teraction with a younger sibling E, even a ﬁve year old north-
ern child C is expected to refer to his sibling D (D is C’s
younger and E’s elder sibling) as anh or chi
or ‘elder sister’). E is not supposed to make third-party ref-
erences from his elder siblings’ perspectives.” So C (oldest)
refers to D (middle) as anh “older brother” when talking to
E (youngest)—thereby adopting E’s perspective. E (youngest),
however, must not refer to D (middle) as em “younger sibling”
when talking to C (oldest). The very act of taking another’s
perspective introduces a “second-order indexicality” insofar
as it presupposes a hierarchical relation between speaker and
addressee (Silverstein 2003). The alternatives for address and
self-reference are thus as follows:
Kinship trope Kinship trope
(pronoun) Pronoun—toˆi, “I”; minh, “I”
13. In Agha’s (2007:352) scheme, such uses achieve the mappings
between speech-act participants and referents via the simultaneous op-
eration of a “denotational” sketch and an “interactional” sketch.
310 Current Anthropology Volume 53, Number 3, June 2012
Here are some examples from a radio interview of two mu-
sicians who are older brother (Pha
c Chaˆu)and younger
m Kha´nh Linh). The interviewer, a radio person-
ality, addresses Pham Kha´nh Linh with her name (example
) and Pham Ngoc Chaˆu with anh (literally, “older
(7) NAME: Kha´nh Linh teˆn d¯yd¯la`Pha
m Kha´nh Linh
Kha´nh Linh name full is Pham Kha´ nh Linh
“My full name is PhamKha´nh Linh”
(Literally, “Kha´nh Linh’s full name is Pham
(8) PNOUN Co`n d¯ u ng thı` mı`nh thı´c h nh t la` u ng nu
`´ ´ ´
And for drinks, comp 1S like best is drinking wa-
ters grapefruit force
“And as for drinks, it’s grapefruit juice that I like
(9) PNOUN: toˆi teˆn la` Pha
1S name is PhamNgoc Chaˆu
“My name is PhamNgocChaˆu.”
(10) Ø: Ø r t thı´ch h oa
Really like ﬂowers.
“I really like ﬂowers.”
(11) NAME: Kha´nh Linh vu`’abola`ho
name just said is little chubby, so conj name
chose clothing Q?
“You just said that you are a little chubby, so how
do choose what to wear?”
(Literally, “Kha´nh Linh said that [you] are a little
chubby, so how does Kha´nh Linh choose what
(12) KTERM: du
’aca`m mmu id¯u´ng khoˆng anh?
Melon eggplant shrimp salt right not older-
“Like, pickled, salty eggplant and so on right?”
(Literally, “Like, pickled, salty eggplant and so on
right older brother?”)
Again, the implications of this for our argument should be
clear. In addressing others and in self-reference, a speaker is
obliged to attend to the relative age or rank of theparticipants
(construed on a model of kinship). Speaking the language
involves constant attention to this aspect of the interactive
context—that is, the facts of how speaker and addressee stand
with respect to each other—and, moreover, requires that some
construal of those facts be included in the utterance. Indeed,
where it is not obvious from physical appearance, personal
introductions frequently include inquiries about the other’s
age. This information is essential to the proper selection of
terms for address and self-reference in all subsequent inter-
There is no possibility for opting out of the system. Al-
though a speaker can employ a personal name in both address
and self-reference and a pronoun in self-reference (as well as
Ø in some contexts), use of these forms is considered ap-
propriate only within highly constrained circumstances. Use
of a personal name, for instance, in self-reference or in address
where the addressee is signiﬁcantly older or in some other
sense deserving of respect may be found “impudent” or
“fresh”—hn. Indeed as Luong shows, even with juniors, use
of a pronoun instead of a kin term is taken to indicate dis-
approbation (see also Agha 2007:357 on this point).
The case studies of Mayan deixis and Vietnamese address
well illustrate the point Silverstein introduced in “Shifters,
Linguistic Categories, and Cultural Description” (Silverstein
1976; see also Jakobson 1971 ): there are signiﬁcant
differences across languages in terms of the indexical signs
that, by their very use, make explicit and overt the contextual
parameters of the speech event. It follows that in speaking,
people are producing contexts that are at least partially shaped
by the language they happen to use. This, then, entails a
second version of the linguistic relativity argument, in which
the particular lexicogrammatical features of a given language
shape the cultural contexts produced in speaking it.
Social Interaction: The Progressive Realization
of Understanding in Action
Our third locus for linguistic relativity is social interaction,
not just in a general sense that would encompass much of
what we have been discussing so far but in the speciﬁc sense
of the sequences of interlocking actions that people carry out
by using language in social settings (Drew 2004; Goodwin
2002, 2006; Goodwin and Heritage 1990; Heritage 1984;
Schegloff 1968, 1996a, 2006, 2007). Social action is about
doing things, where this “doing” involves other people. Lan-
guage is the central tool. We use it to get other people to do
things for us, to help others or inform them of things, to
share experience with them, to afﬁliate with them, or indeed
to disafﬁliate (Enﬁeld 2006, 2009b; Enﬁeld and Levinson
2006a, 2006b; Tomasello 2008). The perspective that is re-
quired for studying this domain of human activity is an “en-
chronic” one (Enﬁeld 2009a:10, 2011b), that is, a focus on
the move-by-move, normative level of “interactional time” as
a complement to other, more familiar temporal-causal per-
spectives in anthropology and related disciplines (phyloge-
netic, ontogenetic, diachronic, synchronic, etc.).
Research on language in interaction using the methods of
conversation analysis has shown that it is both possible and
necessary to examine language use at the micro level in order
to understand just how these forms of social action are ac-
complished and how intersubjectivity is achieved. In the social
sciences, intersubjectivity—joint or shared understanding be-
tween people—is typically explained in terms of convergent
knowledge of the world. On this view, the world exhibits
objective characteristics, and to the extent that different actors
apply equivalent and valid procedures for generating knowl-
Sidnell and Enﬁeld Language Diversity and Social Action 311
edge of the world, they will converge in their knowledge and
understanding of their circumstances (Heritage 1984:26). A
related solution to the problem of intersubjectivity invokes
the notion of a common culture as the resource through
which “the individual’s grasp of reality is mediated” (Schegloff
Conversation analysts have developed a rather different ac-
count of intersubjectivity. One of their key insights (Sacks
1995; Sacks, Schegloff, and Jefferson 1974) was that ordinary
people exploit the systematic properties of conversation in
reasoning about it “online.” For instance, participants in a
conversation can inspect next turns at talk as evidence for
whether and how their own talk has been understood (Moer-
man and Sacks 1988 ). Displayed misunderstandings
can then prompt the initiation of repair in “third position,”
as in the following example (see Schegloff 1992).
(13) Third-position repair, from Schegloff (1992)
01 Anni: Which one::s are closed, an’ which ones are open.
02 Zebr: Most of ’em. This, this, [this, this ((pointing))
03 Anni:r[I ‘on’t mean on the
04 rshelters, I mean on the roads.
05 Zebr: Oh!
07 Zebr: Closed, those’re the ones you wanna know about,
08 Anni: Mm[hm
09 Zebr: [Broadway. . .
In line 01, Annie asks a question. In the turn at line 02,
Zebrach not only attempts to answer it but by virtue of pro-
ducing a response, Zebrach displays an understanding of An-
nie’s line 01 inquiry. From this response, Annie is able to
surmise that there has been a misunderstanding of her talk
in line 01. It appears from the evidence in line 02 that Zebrach
has made a wrong interpretation of “which ones.” Annie is
able to repair the problem in lines 03–04, and the course of
action underway is then reengaged on the basis of the new
understanding which Annie’s correction provides for. As
Schegloff (1991:158) notes, “The ordinary sequential orga-
nization of conversation thus provides for displays of mutual
understanding and problems therein, one running basis for
the cultivation and grounding of intersubjectivity.”
Consider the following case from the opening of a tele-
phone call between two friends, focusing on the two lines in
boldface (04 and 05):
(14) Deb and Dick
01 Deb: [Hello:?hh
02 Dick: Good morning.p
14. Examples from English, Creole, Finnish, and Lao conversation are
presented using the transcription conventions ﬁrst developed by Gail
Jefferson (see Sidnell 2009a, 2009bfor a relatively complete recent glos-
sary as well as some discussion of the application of these conventions
to languages other than English). It is important to note that in this
system, punctuation marks intonation and not syntax.
03 Deb: pHi:, howareya.
04 Dick: Not too ba:d. Howare you?
05 Deb: I’m ﬁ::ne
06 Dick: Howdit g[o?
07 Deb: [.h Oh: just grea:t,!everybody:st- still here.
08 Dick: Oh really(h)p
09 Deb: pYeah
10 Dick: Oh they stayed. Okay.
11 Deb: Yea:h
Dick asks “How are you?” and Deb responds with “I’m ﬁne.”
This simple exchange seems almost vacuous, but in fact it
tells us a lot. Deb’s response in line 05 to Dick’s “How are
you?” in line 04 displays a range of basic understandings of
that turn (Schegloff 1992): by starting to talk at this moment
(and not earlier), Deb shows an understanding that Dick’s
turn was possibly ﬁnished (see Sacks, Schegloff, and Jefferson
1974). By producing an answer, Deb shows that she under-
stands the previous turn to be a question. By answering with
a description of her personal state, Deb shows that she un-
derstands the turn to be a wh- question (involving a question
word like “who,” “where,” “how,” etc.) rather than a yes-no
interrogative. By responding with “ﬁne” rather than “terrible,”
or “fantastic,” Deb shows an understanding of what this ques-
tion is doing in this environment (a routine opening inquiry,
not to be taken literally, etc.). On this view, the use of language
and other forms of communicative behavior in social inter-
action centrally involves the production and recognition of
social action. So when Deb says “I’m ﬁne” in response to
Dick’s “How are you?” she shows that she understood Dick’s
turn to be one of those personal-state inquiries that does not
call for a completely “honest” response but rather an indi-
cation of whether she has any particular bad or good news
to tell Dick (see Jefferson 1980; Sacks 1975; Schegloff 1986).
We see, then, that the turn-by-turn organization of talk
provides for a continuously updated context of intersubjective
understanding, accomplished en passant in the course of other
activities. These publicly displayed understandings are pro-
visional and contingent and thus susceptible to being found
wanting, problematic, partial, or simply incorrect. A ﬁrst
speaker who ﬁnds the understanding displayed by a second
speaker’s turn inadequate has recourse to a mechanism for
correcting it (an organized set of practices of repair; Schegloff
1992; Schegloff, Jefferson, and Sacks 1977).
In these examples, a participant used the available resources
of the English language (potentially along with other forms
of behavior, such as hand gesture) to compose a turn that a
recipient was able to recognize as accomplishing some par-
ticular action. Does the fact that the words were spoken in
English and not in another language have any bearing on how
that action is accomplished? Given the well-documented facts
of signiﬁcant if not radical language diversity, we should ex-
pect the answer to be yes. We should expect that language
diversity has consequences for the constitution of action
One view might be that the structural characteristics of a
312 Current Anthropology Volume 53, Number 3, June 2012
language are inconsequential in this regard: the same actions
get done, in the same ways, regardless of the language used.
Another might be that the available repertoires of social ac-
tions are entirely incommensurate across languages (cf. Zin-
ken and Ogierman 2011; J. Zinken, unpublished manuscript).
Between these extremes lies the position that we want to
defend: some social actions are more readily carried out, or
are carried out in speciﬁc ways, by speakers of a given language
by virtue of the lexicogrammatical properties speciﬁc to that
language. Moreover, while “the same” action in a functionally
general sense (e.g., request, complaint, agreement) may be
possible in different languages, in reality these actions may
differ in speciﬁable and signiﬁcant ways across the languages.
Because an action must be done in a different way, it may
have rather different implications for subsequent action
within the same sequence.
This idea suggests a new program of research. Our goal
here is to explore the direction such a program might take.
If we begin with an illustrative, functionally deﬁned target
action type that we might expect speakers of any language to
want to carry out in social interaction, we can then compare
the speciﬁc lexicosyntactic resources that languages make
available as tools or vehicles for carrying out this type of
action. Do the differences between these linguistic vehicles
for action correspond to differences in the speciﬁc nature of
that action in the case of each language?
We explore this question by focusing on a type of action
that we refer to technically as an “epistemically authoritative
second-position assessment.” By “assessment,” we mean the
use of an evaluative expression (such as “She’s a swell gal”)
to express a person’s stance toward someone or something,
often in the grammatical form of an assertion. (Note that in
the examples below, we also widen the scope beyond “as-
sessments” to assertions more generally; see note 23.) Such
stance taking is an important device for building, maintaining,
and adjusting the afﬁliative links that structure our social
networks. When someone makes such an evaluation in “ﬁrst
position” in a conversation—that is, without being prompted
to do so by another speaker’s prior assessment—this is often
followed in conversation by a similar assessment by a second
speaker (thus, in second position) as a way for the second
speaker to align (or not) with the ﬁrst in stance.
So a second-position assessment is a person’s statement of
subjective evaluation that immediately follows, and thus ap-
15. One property of this second-position assessment is that it may
appear to be prompted by the assessment just prior, implying that the
second speaker might not have thought to make this evaluation on her
own. The reasons for this cannot be detailed here and indeed have only
recently been explored (see, e.g., Enﬁeld 2011b). What we can say is that
there is a pervasive preference for agreement in conversation (Sacks 1987).
Therefore, when someone agrees with a prior assessment, they are often
vulnerable to being heard as just going along with what the other said.
This structurally introduced bias partially explains the use and availability
of pragmatically marked agreement forms such as “of course” in English
(see Stivers 2011).
pears to be occasioned by, and agreeing with, anotherperson’s
evaluation of the same thing. This is a common pattern in
conversation, and one might assume that it is unlikely to be
particularly fraught. But as recent research on English has
shown (Drew 1991; Heritage and Raymond 2005; Stivers
2005; see Stivers, Mondada, and Steensig 2011 and Hayano
2012 for similar work on other languages), even while people
are fully agreeing on the sentiment being expressed (e.g.,“That
child is a handful”), they are careful to monitor and explicitly
acknowledge who has primary rights or greater authority to
make such an evaluation. In an illustrative case from Heritage
and Raymond (2005; also Raymond and Heritage 2006), two
elderly women talking on the phone both voice the opinion
that James, the grandchild of one of the women, is a mis-
chievous boy, but they tussle—in subtle ways—over who has
the primary epistemic authority to make this evaluation. The
notion of relative epistemic authority is captured by Heritage
and Raymond with the simple notation “K-plus” (K⫹) for
the state of “knowing better than the other,” as opposed to
“K-minus.” In their example, the woman who is the grand-
mother of the boy James has the primary epistemic authority
to make this assessment.
As Heritage and Raymond show,
a sequence of ﬁrst assessment followed by second assessment
is harmonic when the second speaker is also the one who has
less authority to know (“K-minus”). Correspondingly, they
show that when the sequence is disharmonic—that is, when
the speaker who produces the agreeing evaluation in second
position is the one who knows, or should know, better—
speakers will go out of their way to redress a ﬁrst assessment’s
implied claim to epistemic priority.
They do so through
various practices of speaking: a less knowing speaker may
manage the disharmony by downgrading their evaluation if
it is in ﬁrst position (e.g., with “I think”), while a more
knowing speaker may upgrade their evaluation if it is in sec-
ond position (e.g., through the use of dedicated preface forms
such as “oh”; see Heritage and Raymond 2005).
16. A complication is that one’s “ofﬁcial status” may be invoked as a
basis for authority, even where this does not align with the reality (e.g.,
a nanny who spends much more time with a boy than his mother does
may know the boy better but may defer to the mother on certain claims
anyway; see Enﬁeld 2011bfor discussion).
17. “Better” in this context can have to do with rights to know, degree
of access, how recently the information was acquired, and so on.
18. Heritage and Raymond (2005) discuss four practices in English
for accomplishing K⫹second assessments. Two of these upgrade by
suggesting that the position was already held or settled. The other two
practices involve usurping the “ﬁrstness” of a previous assessment. Al-
though the authors describe these multiple practices for accomplishing
K⫹second assessments, they do not discuss their functional differenti-
ation in terms of either distribution or what we have described as col-
lateral effects. We may suppose that these different options have different
affordances, and thus it might be that the differences we claim between
different languages are in fact contained within any particular language.
However, because languages differ signiﬁcantly at all levels of structure,
the range of choices in one language cannot be exactly the same as that
in any other. Moreover, where we have found alternate practices in a
given language (for instance, “I know” vs. “if” prefaces in Caribbean
Sidnell and Enﬁeld Language Diversity and Social Action 313
here is in the speciﬁc format of marked practices of speaking
that manage the disharmony of a second-position assessment
being made by a speaker with higher authority to make the
assessment. We will refer to this balancing act—agreeing with
what someone has just said while signaling that one has
greater authority to have said it—as a “K⫹second assess-
ment,” or “K⫹2A.” Think of it as a two-part task: (1) you
want to agree with what was just asserted by the other person,
but (2) you want to signal that you are in a better or more
rightful position than that other person to assert it. As we
see below, the grammatical resources for this practice are
different across languages, and these differences affect the
nature of the action, so that it cannot be done in exactly the
same way across the languages. Because each language’s strat-
egy draws on lexicogrammatical resources that are used for
other functions as well, this introduces “collateral effects” on
how the action is done, as we explore below.
Before we present the cases, let us clarify what we mean
by collateral effects. The notion of collateral effects is of central
importance to our argument, and we believe that it has special
promise for new research in the relation between language,
cognition, and action. Here is how a collateral effect arises.
First, one has a certain end or goal: one wants to do some-
thing. Second, in order to achieve that end or goal, one must
select a means. Third, the means that one selects will nec-
essarily have a certain structure: not only will some elements
of those means be directly responsible for bringing about the
desired ends but these elements will co-occur, often in rela-
tions of dependency, with other features of the structure as
well. Finally, these co-occurring structural features will intro-
duce effects that were not necessarily selected for. These are
collateral effects: side effects of something that was selected
as a means to a required end.
Take a simple example. Suppose you are a student pho-
tojournalist and your instructor gives you the task to “meet
with a real person and succinctly capture his or her person-
ality.” You are given a choice to submit your work in the
form of either a written paragraph of prose or an untitled
photograph. Either of these is a means to solve the task, but
their different affordances introduce different collateral ef-
fects. A collateral effect of using a photograph is that it would
be virtually impossible to avoid revealing the person’s physical
features and thereby things like their age, gender, and state
of health. By contrast, the affordances of prose would readily
allow the writer to leave those aspects of the person unre-
As an example from the realm of symbolic systems, con-
sider expressive differences between the modalities of spoken
language and the hand gestures that accompany speech. Imag-
ine that your expressive goal is to describe a motion event,
say, “He left the room.” If your selected means are in spoken
Creole, or verb-subject [VS] vs. subject-verb [SV] word orders in
Finnish), they are functionally distinct, providing speakers of these lan-
guages with markedly different resources for action.
language only, in the same way that you can describesomeone
in prose without revealing their physical features, you can
verbally describe this scene without making any information
available as to the direction or speed of the event being de-
scribed (as in the English “He left the room”). But if one
chooses to depict this event using hand gestures, one is nec-
essarily showing the motion as having happened at a certain
speed and in a certain direction (regardless of whether one
wanted to show this or whether an onlooker interprets that
speed and direction to be part of what you intend to say).
Of course, we often select cospeech gesture precisely so that
we can exploit these affordances of the modality. But consider
sign language of the deaf, in which one will most heavily rely
on manual means for linguistic expression and not as an
alternative to the vocal channel. When the manual-spatial
modality is used to express motion iconically in sign/gesture
space, not just the fact of motion but, unlike in the vocal
modality, other information about that motion is necessarily
expressed as well. These collateral effects are a product, or
by-product, of the selection of means to ends.
We now turn to our comparison of these kinds of effects
in the pragmatic realm of social action.
Caribbean English Creole: If-Prefaced Repeats
Our ﬁrst example comes from the Caribbean English Creole
spoken on the island of Bequia, St. Vincent (from research
by Sidnell, e.g., 2009c). The action of K⫹second assessment
is routinely done in this language by prefacing a repeat of a
prior speaker’s talk with “if.” First, let us describe the more
common function of the practice of “if-prefacing” in the lan-
In most varieties of English, one standard way of forming
polar questions (i.e., “yes-or-no” questions) is to invert the
ordering of subject and auxiliary verb in simple declarative
constructions (Quirk et al. 1985): “You’re going for a nap”
becomes “Are you going for a nap?” Such inverted syntax can
be preserved in repeats that initiate repair, so that, for instance,
repair is initiated with forms such as “Am I going?” or “Am
I going for a what?” In the creoles of the Caribbean, there is
no auxiliary-subject inversion in polar questions. Indeed,
there is no syntactic category of auxiliary verb for such an
inversion to operate on (see Winford 1993). Instead, in these
varieties a turn’s status as a question is constituted through
a range of features of design and context. None of these
features (intonation and prosody, directed gaze) can be trans-
ferred (unproblematically) to a turn that other-initiates repair
in order to show that the turn being targeted was understood
to be a polar question (see Sidnell 2009cfor further discus-
sion). This is where if-prefacing comes in.
In their basic canonical use, then, if-prefaced repeats are
used to initiate repair of a prior turn that is formatted as a
polar question. Pat’s turn at line 03 of example (15) is an
instance. Here Pat and Benson are sitting side-by-side in the
yard that adjoins Benson’s small house. It is a week after
314 Current Anthropology Volume 53, Number 3, June 2012
Carnival, and Pat has stopped by for a visit with Benson’s
(15) #187_Q2 qt 51:50
01 Benson: yu biin hii fu kanival (.)Pat?
were you here for Carnival Pat?
03 Pat:rif mi bin wa?
if I was what?
04 Benson: Bekwe fu kanival?
05 Pat: yeah:
When Benson asks Pat in line 01 if she was in Bequia for
Carnival, Pat responds by initiating repair with if mi bin wa?
“If I was what?”—an if-prefaced repeat. The preframing mi
bin, “I was,” combined with the question word wa, “what,”
isolates hii fu kanival, “here for Carnival,” as the trouble
source to be repaired. By prefacing the turn with if, Pat also
shows that she heard the turn addressed to her as a polar
question. At line 04, when Benson repairs the reference, he
preserves the status of his turn as a question by producing it
with rising intonation. At line 05, Pat answers the question
with a “yeah.” The function of the if-preface in line 03, then,
is to show that the one initiating repair heard that the turn
containing the trouble source was produced as a polar ques-
tion. Thus, in their basic, canonical use, if-prefaced repeats
initiate repair of a prior question turn and in so doing begin
an insertion sequence that breaks the contiguity of the ﬁrst
and second parts of an adjacency pair (Schegloff 2007).
If-prefaced repeats are also deployed in a quite different
sequential context—speciﬁcally, in response to assessments.
In this environment, if-prefaced repeats are second assess-
ments that agree with a prior. Example (16) provides an il-
lustration. Here, Shanka and Kitana are sitting in their family
yard with three young children close by. At line 01, Kiki directs
one of these children—Zaria (who is Shanka’s cousin and
Kiki’s niece)—to move and, in close succession, to “pick up
that thing,” an object on the ground in front of her. Kiki
completes the turn with a question about the whereabouts of
Zaria’s cousin Roxanne. Zaria responds only to the ﬁnal part
of the turn, apparently pointing to where Roxanne is. Kiki
then directs her attention to the slightly older girl Naksin. As
Kiki produces this turn, Zaria runs away. This occasions
Shanka’s evaluative assessment of Zaria at line 05—Wailnes
Zaria a kom wid, “Wildness Zaria comes with.” At line 06,
Kitana responds with if Zaria wail?, “if Zaria is wild.”
(16) #134_Q1 qt 25:38
01 Kiki: Zaria muv from de.ptek a ting. wapa Rakzan.
Zaria move from there. Take that thing. Where is Roxanne.
02 (3.0) ((Zaria apparently points to where Roxanne is))
03 Kiki: Naksin lii shii:. tek da- an ting- Jak ting.
Naksin leave her, take that and thing, Jack’s thing
04 (2.0) ((Zaria runs away))
05 Sh: Wailnes Zaria a kom wid.
Wildness Zaria comes with
06 Kiki:rif Zaria wail?
if Zaria is wild
08 Naksin kom.
Naksin come here
We can see, then, that this assessment sequence is occasioned
by a complex set of visible behaviors and witnessable actions:
the failure to comply with the directives, the ﬂailing hand
gestures, the running away. Note that, unlike example (15),
here the if-prefaced repeat does not elicit any response from
the recipients. Indeed, it seems to close the assessment se-
quence, as the talk turns to other matters.
Consider next example (17). This begins with Donna call-
ing to her nephew (who is off camera). Although he appears
to respond, he does not comply with the request to “come
here.” After Kiki beckons the same boy again, Benson turns
to Donna and remarks, “He’s rude you know.” This initiates
a string of assessments culminating in an if-prefaced repeat.
(17) #139_Q1 qt 39:25
01 Donna: Gusnel kom bai hee
Gushnell come over here.
02 (Gushnell): (for yu)
03 Kiki: (Gusnel)kom he.
04 Benson: ⬚hii ruud yuno⬚((Benson turns behind him
he’s rude you know to look at Donna))
05 Donna: ai noo hi ruud
I know he’s rude
07 Benson: riil ruud. ((returns gaze in front of
real rude him))
09 Ezekiel: huu ruud.
10 Benson: da boi [de.
That boy there.
11 Donna:r[if hi ruud?
if he’s rude?
The if-prefaced turn in line 11 once again closes this extended
sequence of assessments. With it, Donna seems to have the
last word, and the talk turns to other concerns. Note that in
both this and the previous example, the original ﬁrst-position
assessment is a vehicle for complaining about a nonpartici-
pant third party.
As pointed out above, in their basic interactional environ-
ment, if-prefaced partial repeats initiate repair on a polar
question. In other words, if-prefaced partial repeats convey
that their speaker has heard a previous turn containing the
trouble source to be a polar question. In these last two ex-
amples ( and ), the practice is used to treat a prior
assessment as if it were a polar question. It is important to
recognize in this respect that in examples (16) and (17), the
turn to which the if-prefaced assessment responds is not, in
fact, a question. Moreover, in examples (16) and (17), the if-
Sidnell and Enﬁeld Language Diversity and Social Action 315
prefaced turn does not initiate a repair/insertion sequence, as
it does in example (15). Rather, with these if-prefaced second
assessments, second assessors are doing agreement. One piece
of evidence for this is that like other agreements (and pre-
ferred actions more generally), these turns are closing-impli-
cative, meaning that they help to bring topics or sequences
to a close (but see our discussion of VS-formatted responses
in Finnish below). By contrast, dispreferred actions and dis-
agreements in particular tend to be sequence elaborative and
to engender more talk on the same topic. So note that in
example (16), after the assessment sequence at line 05–06 the
talk turns to other matters (Kiki beckons Naksin) and that
in example (17), a long string of assessments concludes with
an if-prefaced repeat, at which point the participants again
turn to other matters.
But these if-prefaced turns go beyond just agreeing. By
responding to the ﬁrst assessment as if it were a polar question,
the if-prefaced turn treats a ﬁrst assessment as a question and
thus as epistemically downgraded relative to a declaratively
formatted assertion (see Heritage and Raymond, forthcoming,
on the notion of an “epistemic gradient”). By considering the
context in which the practice is used, we ﬁnd evidence in
support of this analysis. In example (16), for instance, the
ﬁrst assessment is produced by Shanka, who is Zaria’s cousin,
and the second by Zaria’s aunt (Kiki), who is partially re-
sponsible for her. The assessment here is a complaint, and
thus the if-prefaced format of the second assessment may be
selected to deal with a situation in which Kiki feels she needs
to reassert her greater rights to evaluate the child (see Ray-
mond and Heritage 2006). In example (17), evidence of the
participants’ orientation to the matter of differential epistemic
rights is found both in the prior talk and in the social relations
by which they are connected to one another and to theperson
being assessed. Here, Benson is assessing Donna’s nephew.
The design of the initial exchange is sensitive to Donna’s
greater epistemic rights. Speciﬁcally, Benson produces his ﬁrst
assessment as a question (line 04: ⬚hii ruud yuno⬚, “He’s rude
you know”), in this way inviting Donna to conﬁrm it, which
she does in a particularly explicit way (line 05: ai noo hi ruud,
“I know he’s rude”), not only conﬁrming but explicitly re-
ferring to her claimed knowledge state.
The sequence con-
tinues with Ezekiel initiating repair (huu ruud, “Who is
rude?”). After Benson repairs the reference with a demon-
strative referring expression, Donna responds to the initial
assessment again, now with an if-prefaced turn. Here again,
then, relative rights to assess are at issue: this is, after all,
Donna’s nephew that Benson and Ezekiel (a family friend)
are assessing. The participants’ orientation to the matter of
differential rights to assess, here grounded in different social
relations to the person, is made explicit through the design
of the initial assessments. In using the if-prefaced second as-
19. This is an alternative format for accomplishing a K⫹second as-
sessment in this variety, one that would appear to engender quitedifferent
collateral effects (e.g., it is not closing implicative).
sessment at line 11, Donna is pushing these already recognized
rights to their limit—claiming, in effect, the last word on the
matter of whether the child is rude.
If-prefaced second assessments work the way they do be-
cause they treat a prior assessment as if it were a question.
If-prefacing takes its sense and import in this environment
from its canonical use in the other-initiation of repair of a
prior polar question. Its use in K⫹second assessments is
arguably derivative of a more basic use in initiating repair,
and this in turn appears to be a reﬂex of grammar in languages
that do not use syntactic inversion to form polar questions.
The example of if-prefacing thus shows how speciﬁc gram-
matical patterns can have consequences for action in talk-in-
interaction, as a result of the features of a lexicogrammatical
vehicle for action that are imported as collateral effects of
doing that action. Speciﬁcally, in initiating repair of polar
questions (and in reporting them; see Sidnell 2009c), second
speakers may preface the turn by “if.” Once established as an
“interrogative marker,” this item “if” may be adapted to other
contexts, thus opening up language-speciﬁc possibilities of
social action. As we have seen above, with an if-prefaced
second assessment, a second speaker not only agrees with a
ﬁrst assessment but also claims greater epistemic rights to
assess the matter. A further collateral effect has to do with
how this practice shapes the context for subsequent talk. Spe-
ciﬁcally, if-prefaced second assessments are closing implicative
to such an extent that their speakers can be heard as having
the last word (or trying to have it; see Sidnell 2009cfor a case
in which the attempt fails). So in this case, we see at least
three interactional functions fused in a single practice: (1)
agree with prior assessment, (2) claim epistemic priority rel-
ative to the ﬁrst-assessment speaker, (3) move to close topic.
As we shall see next, while there are similar practices in other
languages, these are not identical.
Finnish: Word-Order Variation
Our second example of the consequences of language-speciﬁc
grammatical patterns for social actions comes from research
on Finnish by Hakulinen and Sorjonen (2009; Sorjonen and
Hakulinen 2009). These authors discuss a range of alternative
second-assessment formats, all of which are used to agree with
a ﬁrst assessment.
Two typological features of Finnish turn out to be relevant
in the formulation of utterance formats as means of agree-
ment to assessments. First, Finnish is a language that has,
from a grammatical point of view, “free word order” (cf.
Vilkuna 1989). This means that for example verb initiality
can be deployed for a number of discourse purposes, one
of them being the responding to questions, assessments and
negative assertions. Secondly, in Finnish, a fully grammatical
clause can be formed without an overt subject: with an
anaphoric zero, a response is tied to the utterance of the
prior speaker. (Hakulinen and Sorjonen 2009:149)
316 Current Anthropology Volume 53, Number 3, June 2012
The result is that Finnish speakers have ﬁve distinct ways
of repeating some portion of a ﬁrst assessment in order to
agree with it.
For the purposes of illustrating our argument,
we focus on just two of these responses, described by Sorjonen
and Hakulinen “VS” and “SV.” These alternatives are shown
in the following example (adapted from Sorjonen and Hak-
First assessment: Se mekko on hieno
That dress is great
Response: se on SV
on se VS
According to Sorjonen and Hakulinen, the ﬁrst of these re-
sponses conveys the independence of the second speaker’s
stance—that the second speaker, while agreeing with the prop-
osition, held the view being expressed independently of its
being articulated by the ﬁrst-assessment speaker.
response conveys that the second-assessment speaker agrees
with the ﬁrst but views the matter from a different perspective.
That difference may or may not be explicated in the ensuing
talk. The following example illustrates the second, VS format.
Here, the participants are talking on the phone and discussing
the weather. In the preceding talk, each has described the
beautiful weather in their present location (L is at home, and
A is at her cottage in a place where L also has a cottage).
Where the transcript begins, A transitions from talk about
the weather conditions of the day to “autumn” generally.
(19) Sorjonen and Hakulinen 2009
09 A: [Nyt kelpais olla]k[i. Mut kylla¨ sie]lon
[now worth.would be.cli but sure there is
[Now it would be something. But it is really
10 L: [Ai() ]
11 A: ihanaa heti kun ei sada.
lovely immediately when neg rain
lovely there as soon as it’s not raining.
12 L: FNii,[Joo,
13 A:r[.hh Kyl se on: syksy on niin mahdottoman
[prt it is: autumn is so impossibly
[.hh It really is: the autumn is so extremely
L: [1Kylla¨ma¨vi-ei viikonloppuna menen!.hhh ma¨ meen
20. The authors describe these responses as forming a “paradigm,”
labeling them as follows (where V is verb and S is subject): V, V⫹V,
V⫹S, S⫹V, V ⫹particle. And of course, there are additional ways of
agreeing that do not involve repetition.
21. As the authors note, this is similar to the use of “oh” in English
in second assessments (see Heritage 2002a).
[1sure I we- neg weekend go.1 I go.1
[1I’m surely wee- at the weekend I’m going!.hhh I’m
18 kans t- ka¨a¨nta¨a¨Fmaat ja .hh laittamaan kuntoon varmuuden
also turn lands and put shape.ill safety’s
going to dig the land over and .hh put everything in
shape to be
19 vuoks kaikki jos (.)jos sitte ei tuu ena¨a¨⬚mennyks⬚
sake everything if if then Ø neg comes anymore going
on the safe side if (.) if one doesn’t get to go there ⬚any-
So, as Sorjonen and Hakulinen explain, while the participants
here clearly agree on the beauty of autumn, they are never-
theless positioned rather differently with respect to the prop-
osition “the autumn is beautiful.” A is at this moment en-
joying the beauty of the Finnish hinterlands in autumn and
by expressing this as a bald, timeless statement of fact (the
autumn is so extremely beautiful) can perhaps be heard as
anticipating future occasions on which these same conditions
will be enjoyed. L, on the other hand, is, it turns out, preparing
to sell her cottage. After she agrees with the assessment “the
autumn is so extremely beautiful” using a VS-formatted re-
sponse, she goes on to explain that she plans to visit her
cottage, potentially for the last time, the next weekend to “put
everything in shape.” This, as Sorjonen and Hakulinen note,
is the beginning of a troubles-telling and thus conveys quite
explicitly the difference in perspective that the VS-formatted
second assessment adumbrated. Key to our argument here is
the authors’ observation about this and other examples like
it: namely, because the VS format implies a difference of
perspective, it can be topically elaborative. That is to say, use
of this format can establish the relevance of an unpacking of
the different perspective of the second speaker and thus result
in elaboration of the topic.
Consider now the following instance, in which the response
(20) (Field note, spring 2007)
01 A: Se on ihan hirveen hyva¨ opettaja.
it is quite terribly good teacher
He is just an extremely good teacher.
02 B:rSe on.
Sorjonen and Hakulinen (2009:288) explain:
By responding with the format se on B both asserts agree-
ment with A, and implies the independence of her stance.
In the prior talk, B was the one who was telling the others
about the speciﬁc event, portraying the success of the col-
league’s teaching. Now that A presents her assessment, B’s
turn can be heard as agreeing but simultaneously conﬁrming
The following example provides a more complex case but
one for which we have a fuller sense of the context, including
Sidnell and Enﬁeld Language Diversity and Social Action 317
the ensuing talk, which provides the evidential basis for our
claims about the character of the action being performed.
(21) (Kotus, T1208: 61, eastern Finland, hairdresser’s)
01 C: Joo.
02 (33.7) H cutting Client’s hair
03 H:r.mt Voi mahoto mite o itsepa¨ine hius.
oh impossible how is obstinate hair
.tch Oh my god what obstinate hair.
04 C: Mm-m.
06 C:r.nf Ne on.
08 H: Mite sie sita¨ aina kotona ite laitat.
how you it-par always home.ess yourself make.2
How do you set it at home by yourself.
09 C: No geelilla¨ha¨n sita¨ pitta¨a¨⬚mh⬚,(0.7) muottoilla.
well gel.ade.cli it-par Ø must shape
Well you have to use gel ⬚mh⬚, (0.7) to shape it.
Here the ﬁrst assessment in line 03, “mite o itsepa¨ine hius,”
has a singular referent and employs a lexical referring ex-
pression rather than a pro-term. The response in line 06
substitutes a plural pro-term for hius, “hair,” but is never-
theless formatted as SV. Note now that the independence of
C’s stance—that she felt this way before the current occasion,
independently of what has just been said—is presumed by
the ensuing question from H (“How do you set it at home
by yourself?”). This very question presumes that the client
has encountered on previous occasions the obstinacy that the
stylist has remarked upon and the client conﬁrmed.
So in the Finnish case, there are at least two distinct formats
for agreeing with a prior assessment while at the same time
asserting independent epistemic access (K⫹)fromsecond
position. The availability of word-order alternations (among
a number of other grammatical features; see Hakulinen and
Sorjonen 2009) makes possible a distinction between [K⫹]
as expressing independent access and [K⫹] as expressing a
different perspective. And this has the collateral consequence
that the use of the VS format can be topically elaborative
rather than closing implicative (in contrast to the Creole case
we have just seen; cf. also Lao, below).
In both the Creole and the Finnish cases, the second, agree-
ing, K⫹assessment involves repeating a portion of the prior
talk. But the languages differ substantially in the way the
repeated portion is elaborated by the available lexicogram-
matical resources of the language. In the Creole, speakers use
a form the basic semantics of which derive from its use in
the other-initiation of repair (i.e., if-) to cast the prior as-
sessment as a question, thereby retrospectively transforming
the other speaker’s ﬁrst assessment by suggesting that it was
said with uncertainty. In Finnish, relatively free word order
makes possible an alternation between VS- and SV-formatted
repeats. This provides for a distinction within the broader
category of K⫹responses, such that the VS format implies
a different perspective. By comparing the cases of Creole and
Finnish, we see that they both provide speakers with a way
to do K⫹second assessments, but the linguistic resources
that are picked up and used as tools for this speciﬁc action
are associated in the two languages with other, nonequivalent
functions. This nonequivalence gives rise to different collateral
effects on the speciﬁc nature of the action in the two lan-
guages. We now turn to our third case, Lao.
Lao: Perfective Particle le`q1
Lao is a Southwestern Tai language spoken in Laos, Thailand,
and Cambodia (Enﬁeld 2007a). An important grammatical
feature of this language, as of many other languages of the
mainland Southeast Asia area (Enﬁeld 2005, 2011a), is the
frequent (though not obligatory) use of ﬁnal particles. Lao
has a large inventory of illocutionary particles that typically
occur in ﬁnal position relative to a turn-constructional unit
(i.e., turn-ﬁnally, attached to either a clause or a freestanding
noun phrase; Enﬁeld 2007a, chap. 4). These are words that
occur at the ends of utterances (whether those utterances are
full “sentences” or fragments), with the function of indicating
certain grammatical meanings, such as temporal features of
an event (tense, aspect), or shades of meaning relating to the
attitudes and expectations of the speaker, often in relation to
the addressee. There are three broad, functionally deﬁned
categories of Lao ﬁnal particles: interrogative, imperative, and
factive. Loosely speaking, the factive particles are appended
to assertions, forming a set of contrasting semantic speciﬁ-
cations concerning the epistemic status of a proposition, given
aspects of the speech event. For example, one may take a
simple proposition such as man2 phe`e`ng2, “It’s expensive,”
and add different particles for different effects: with the par-
ticle deˆj2, a speaker makes a claim that the assertion is news
to the addressee (man2 phe`e`ng2 deˆj2, “It’s expensive you
know”); with the particle sam4, a speaker makes a claim that
the proposition is unexpected or surprising, given the context
(man2 phe`e`ng2 sam4, “But as it turns out, it’s expensive”);
or with the particle le`q1, a speaker makes a claim that the
assertion is already or independently known to be the case
(man2 phe`e`ng2 le`q1, “It is indeed expensive”).
One way in which speakers of Lao manage the action of
making a K⫹second assessment is to append the particle
le`q1 to the repeated evaluation or assertion. This particle,
exempliﬁed in the previous paragraph, is termed a “factive
perfective.” It is elsewhere used to signify completion of an
action or event, a meaning arising from its association with
a perfective verbal marker le`e`w4 (which is also a full verb
meaning “to ﬁnish”; Enﬁeld 2007a:180ff.). The connection
between the two meanings “completedness of an event” and
“prior establishment of the truth of an assertion” should be
obvious: generally, if I can assert that a narrated event is
completed, I convey in the speech event that I have inde-
pendent knowledge of its truth. Syntactically speaking, as a
ﬁnal particle, le`q1 may occur appended to clauses or noun
318 Current Anthropology Volume 53, Number 3, June 2012
phrases (usually demonstratives), as in the following two ex-
(22) bo`ø qu`t2 le`q1
neg lacking fac.prf
Indeed ((it’s)) not lacking.
(23) nan4 le`q1
((It’s)) that ((very)) one.
The communicative function of le`q1 is to signal independent,
prior, or markedly precise knowledge of the assertion. In the
following case (from a video recording of an informal visit
between in-laws in rural Vientiane, Laos), speaker M uses le`q1
to signal this sense of special precision. The sequence begins
in line 2 with speaker K’s question as to whether speaker P
has fully recuperated after an accident. Speaker P responds
with a minimal conﬁrmation (in line 3), and speaker M (P’s
wife) follows this up with an elaboration, ﬁrst stating that P
has been able to walk for 2 days and then clarifying with the
more precise statement that it has in fact been 3 days. This
clariﬁcation (in line 6) is marked with le`q1.
2K khaj1 de`e`1 laaj3 teep5 laø mbo`øp
improve somewhat much very prf pcl
((Your condition has)) improved quite a lot, right?
4M phaa3-ko`ø n˜aang1 daj4 so`o`ng3 mu`u`4
just walk can two day
((He is)) just able to walk, ((since)) two days ((ago)).
6saam3 mu`u`4 niø le`q1
three day this fac.prf
Three days today ((in fact)).
Similarly, in the next case (from a video recording of two
neighbors talking in a rural Vientiane village), speaker S is
talking to speaker K, who has just arrived in the village by
car. Just before the segment supplied here, speaker S has been
asking about another man named Loy (K’s son), and S has
asked whether Loy had also just come with K to the village.
In the ﬁrst lines of the example (lines 15–16), K explains that
when Loy had the chance to get in the car to travel to the
village, he had not done so. In line 18, S asks if this is because
Loy was tied up with work. When K replies that he was not
and that Loy had simply not wanted to come, this information
is marked with the ﬁnal particle le`q1 (line 20). It conveys an
air of ﬁnality to what is being said, in line with the perfective
semantics of le`q1. Accordingly, the appended question in line
24 cannot be heard as anything other than rhetorical, and
indeed the sequence closes here completely, with 20 s of si-
15 K pho`ø-dii3 vaa1 khu` n5 lot1 juu1 han5 bak2-looj3 phen1
right.when say ascend vehicle loc there m-L. 3sgp
16 bo`ø maa2 baat5-niø
neg come pcl
Right when ((we)) said ((let’s)) get in the car there, Loy, he
didn’t come, now.
18 S khaa2 viak4
((He was)) tied up with work ((you mean?)).
20 K kaø bo`ø khaa2 le`q1
topic.link neg stuck fac.prf
((He was)) not tied up!
22 tang4 bo`ø maa2 su`-su`u`1 niø le`q1
intend neg come that’s.all tpc fac.prf
((He)) just didn’t come, that’s all!
24 K siø khaa2 n˜ang3
irr stuck what
What would ((he)) be tied up with?
As these examples show, a central function of le`q1 is to convey
a sense of “ﬁnished,” a function that is clearly traceable to
the source of this word in the verb le`e`w4, meaning “to ﬁnish.”
When this meaning is imported from the realm of the nar-
rated event (i.e., what the utterance is about) into the realm
of the speech event (i.e., the speech-act participants and their
relationship), it comes to mean “there’s nothing more to be
said here now.” It is this component of the ﬁnal particle le`q1
that makes it appropriate as a vehicle for managing the K⫹
second assessments that we have focused on above. We now
turn to examples in which le`q1 is used for carrying out this
type of action, that is, where le`q1 is appended to second-
position assessments (or assertions, to be more accurate with
reference to the Lao examples below) where the speaker is
agreeing, on the one hand, but pushing back against a seeming
claim to authority embodied in a ﬁrst-position assessment,
on the other.
In the following example, two elderly villagers, Mr. Ka and
Mr. P, are having an informal conversation at the village home
of Mr. P. They are both lifelong residents of the rice-farming
plains and nearby forests and waters around the area of Vien-
tiane. The occasion is a visit by Ka (and his wife) to P (and
his wife). The two men do not know each other well, and as
they discuss the local forests and the various herbal medicines
that can be found there, a sense of competitiveness arises
concerning their relative knowledge and expertise. At the be-
ginning of this extract, as P holds a piece of a certain herbal
medicine, he states that some people mistake it for another
kind called haak phang khii,
but in fact he knows that it is
from a plant called kok sii din. In the section of interest to
us here, Ka then begins with an assertion that the herb known
22. In the recording, it is not entirely clear that the speaker says haak
phang khii, as opposed to some other medicine; hence the bracketsaround
the words in the transcript here. This does not detract from the point
being made here.
Sidnell and Enﬁeld Language Diversity and Social Action 319
as haak phang khii is plentiful (or “not lacking,” as the idiom
goes) and is on his way to stating where it is plentiful (line
12), but P begins talking and names the place in line 14:
“Vang Pheˆeˆng Weir.” This is conﬁrmed, or at least accepted,
in line 15 by Ka. Here we are interested in line 16, spoken
by P, which ties back to Ka’s utterance in line 11. Speaker Ka
had used the expression bo`ø qu`t2, “not lacking,” in making
an assertion about the herb known as haak phang khii,and
this expression is picked up in line 16, though this time—
critically—with the addition of the factive perfective particle
le`q1. By adding the particle, P carries out the double-barreled
action we have seen in the above examples: on the one hand,
he is fully agreeing with what his interlocutor has said, while
on the other hand, he is making it explicit that he has a
greater authority to have said it. His “agreeing” utterance in
line 16 therefore has a “conﬁrming” character (though of
course he had not been asked for conﬁrmation).
9P laang2 khon2 khaw3 vaa1 (haak4 phang2 khii5)vaa1 san4
some person 3pl say plant sp. say thus
Some people, they say (it’s haak phang khii), ((they)) say.
10 bo`ø me`e`n1
((That’s)) not so.
11 qanø-nii4 kok2 sii2 din3 [qanø-niø qaø
clf -this plant sp. clf-this pcl
This is kok sii din, this.
12 Ka [haak4 phang2 khii5 kaø bo`ø qu`t2
plant sp. topic.link neg lacking
Hak phang khii is plentiful, at the area of-
14 kaø cang1 vaa1 faaj3 vang2 pheˆeˆng2 faaj3 n˜ang3 qooj4
topic.link so say weir VP weir what intj
Like ((I)) said, Vang Pheˆeˆng Weir, whatever weir, oy.
15 Ka m5
16 P bo`ø qu`t2 le`q1, faaj3 qanø nanø naø
neg lacking fac.prf weir clf that tpc
((It’s)) not lacking, that weir.
A second example of the same pattern is from a conver-
sation between K and S, two men living in neighboringhouses
in the same village. Again, there is a certain competitiveness
in the air, with K and S disagreeing as to whether certain
rural roads around their village are currently navigable. The
condition of rural roads is a pervasive topic of discussion and
monitoring, as they are always changing because of weather
and trafﬁc conditions. In lines 31–33, K is offering evidence
from his own recent experience to support his claim that a
certain route is difﬁcult to navigate because it is too sandy.
S provides what looks like a solution to this in line 34, by
stating that it would be easy to take a route that passesthrough
a village called “Kilometer 40.” K’s next turn (in line 35)
repeats the key information—“enter (at) Kilometer 40”—yet
adds the ﬁnal particle le`q1, as if this were a kind of conﬁr-
mation or more speciﬁc statement of what the prior speaker
had just said (even though it is identical with it). The effect
is to bring about the target action: agreeing, yet claiming
greater authority to make the assertion.
And through the
speciﬁc tool chosen for this action, a collateral effect is that
the sequence is closed down by virtue of the basic semantics
of the particle.
31 K pajø phe`e`1 khaw5 kuu3 paj3 hanø le`q1
go distribute rice 1sg go tpc.dist fac.prf
((When I went)) to distribute rice, I went ((to that place)).
32 cang1 vaa1 man2 peˆn3 thaang2 khiø-saaj2
thus say 3sg be path sand
Therefore, ((that’s why I)) said it was a sand road.
33 [vaang1 hanø
34 S [maa2 phii4 maa2 phii4 lak2 sii4-sip2 phii4 sabaaj3
come dem.prox come dem.prox km 40 dem.prox easy
Come here come here (Km 40 here), easy.
35 K khaw5 lak2 sii1-sip2 niø le`q1
enter km 40 tpc fac.prf
It is indeed at Km 40 that [you] enter.
In these examples, we have shown how Lao speakers con-
ventionally construct a K⫹second assessment by exploiting
a lexicogrammatical resource that is characteristic of one of
the language’s basic organizational properties, ﬁnal particles
(Enﬁeld 2007a, chap. 4). The particular format chosen—the
factive perfective particle le`q1—is well ﬁtted to this function
because of its source in a perfective aspectual meaning, along
the lines of “ﬁnished.” This lends it an air of “shutting down”
a sequence of interaction.
Our comparative case study shows how three very different
languages from three corners of the world provide their speak-
ers with different resources for constructing a common type
of social action in interaction: agreeing with what someone
just said. These actions—saying things and agreeing with what
others say—are crucially implicated in the way members of
our language-using species afﬁliate with one another and thus
display solidarity. These linguistically mediated actions are
central to our sense of having a common outlook with some-
one or some group and hence are central to forming and
maintaining alliances and relationships (Enﬁeld 2009b, 2011b;
Sidnell 2010). We have shown that each of the languages
discussed provides sufﬁcient tools for carrying out the speciﬁc
social action we have characterized as a K⫹second assess-
23. Clearly, this is, in surface terms, an agreement not to an assessment
but rather to an assertion. This does not, however, detract from the
essential function of this format: agreeing while claiming epistemic au-
320 Current Anthropology Volume 53, Number 3, June 2012
Figure 2. Collateral effects of selecting language-speciﬁc formats for carrying out a general type of social action, “K-plus agreeing
second assessment” (K⫹2A). To carry out a certain action, a speaker has no option but to select some language-speciﬁc means for
doing it; these different means introduce different effects in each language, resulting in different ﬁnal outcomes. YNQ pyes-no
question; WO pword order; PCL pparticle.
ment—that is, agreeing with what someone just said while
conveying that one has greater authority to say it—yet because
of the particular tools conventionally chosen for that action,
each language brings the action about in a distinct way. This
is the third sense of linguistic relativity that we are proposing:
differences in language-speciﬁc structures available to differ-
ent speech communities give rise to differences in the ways
that speciﬁc social actions are enchronically effected, thereby
changing how these actions play out in sequences of social
interaction. We have argued that these consequential differ-
ences are the outcome of collateral effects arising from prop-
erties of the linguistic structures that serve as tools for carrying
out the target social action and that are unavoidably intro-
duced when the structure is selected.
To study these linguistically relative collateral effects, there
are two possible points of focus. First, we can focus on nu-
ances of meaning in different language strategies. For example,
we can compare the Creole case, in which the agreement
format treats a prior as if it had been a question, with the
Lao case, in which the particle le`q1 draws on the semantics
of factivity and perfectivity and by this comes to express ﬁ-
nality of the agreement turn in interactional and pragmatic
terms. Such effects can be subtle. But because they are nec-
essarily cumulative and combinatorial and because every so-
cial action is necessarily tooled by the particular semiotic
resources of the language in question, we contend that these
subtle effects will result in substantial differences in the ways
that general patterns of interaction are inﬂected (see also Lev-
inson 2005; Schegloff 2006; Sidnell 2007a, 2007b, 2009a). This
is especially clear in the second possible point of focus for
studying linguistically relative collateral effects, namely, the
consequences of a particular action format or practice for
subsequent talk. This is seen particularly clearly in the com-
parison of Lao and Creole second assessments with Finnish
VS-formatted second assessments. While the former are im-
plicative of sequence closure—indeed, they seem to suggest
insistence on having the last word—the VS-formatted second
assessments of Finnish imply different perspectives and as a
result may be topic elaborating.
To summarize the ﬁndings of our case study, it may help
to visualize the phenomenon of collateral effects in action
(ﬁg. 2). The ﬁgure illustrates how a general action type is
conventionally approached in distinct ways in the three lan-
guages, in terms of the lexicosyntactic resources that each
language provides as vehicles for that action type. Because of
the different properties of the language-speciﬁc vehicles, each
language enables speakers to hit the same broad target but
never quite in the same way: different collateral effects are
In conclusion, the question of linguistic relativity continues
to be debated in current anthropology, attracting as much
interest as ever. Amid continued debate and further studies
concerning the consequences of language diversity for cog-
nition and cultural context, we have tried to formulate a new,
third direction for this work. Our third locus for linguistic
relativity can be found in the enchronic context of social
action as carried out through talk in everyday interaction.
Because such social action is done with the tools that our
languages provide and because these tools are structurally
overdetermined through their rich meanings and multiple
Sidnell and Enﬁeld Language Diversity and Social Action 321
functions, the conventionalized selection of such tools will
have language-speciﬁc collateral effects on the ﬁnal nature of
the action. On this view, the language you speak makes a
difference in the social actions you can perform. The lan-
guage-speciﬁc vehicle or means for an action—even where
that action is a general goal or end that we expect people will
want to pursue in any cultural context—will shape the action
as a function of the structures it introduces. Our case study
suggests ways in which a general target action type can be
tooled in different ways by different languages (and, indeed,
where such tooling cannot be avoided) because of structural
differences in the language-speciﬁc vehicles for conventionally
carrying out that action. By selecting a certain lexicosyntactic
vehicle as a means for achieving social-action ends, speakers
unavoidably introduce associated features, thereby introduc-
ing the collateral effects that we suggest are imported by lim-
itations of lexicosyntactic resources for the construction of
social action through primarily linguistic turns at talk. This
means that differences in language structure are notassociated
only with differences in patterns of thought or cultural con-
text. Differences in language structure lead to linguistically
relative collateral effects, which lead in turn to differences in
our very possibilities for social agency.
This article began as a 2009 draft by Sidnell suggesting a third
form of linguistic relativity in the domain of social action,
sent to Enﬁeld for comments in June 2009. Subsequent dis-
cussions led to signiﬁcant collaboration on development of
the central idea and coauthorship of the ﬁnal version. We
gratefully acknowledge comments and suggestions for im-
provement of earlier drafts from ﬁve anonymous reviewers,
Mark Dingemanse, Paul Drew, Susan Ehrlich, Kobin Ken-
drick, Paul Kockelman, Steve Levinson, Tanya Romaniuk,and
members of the Socio-Cultural Linguistic Working Paper Se-
ries at the University of Toronto, especially Frank Cody, Nais
Dave, Alejandro Paz, Lorna Pitre, Ivan Kalmar, and Michael
Department of Anthropology, University of California, Los Ange-
les, 341 Haines Hall, Los Angeles, California 90095-1553, U.S.A.
(firstname.lastname@example.org). 8 I 12
Jack Sidnell and N. J. Enﬁeld have proposed “a new program
of research” for examining linguistic relativity that looks at
the ways in which language-speciﬁc lexical-grammatical re-
sources are used to accomplish particular types of social ac-
tions in conversation. They have chosen “epistemically au-
thoritative second-position assessments” as a test case. These
are assessments about someone or something through which
a speaker agrees with what was previously asserted by another
speaker and at the same time claims a higher “epistemic au-
thority.” Starting from the observation that languages do the
job of performing these assessments with different linguistic
resources, the authors show that the use of each type of re-
source brings about a different outcome. There are at least
two important assumptions in the project, both of which are
based on prior research by conversation analysts. One is that
conversationalists have a preference for agreement (hence, the
default second assessment is usually in agreement with the
ﬁrst), and the other is that speakers want to make sure that
their superior rights to know about someone or something
are recognized and therefore mark their utterance appropri-
ately (how such “superior rights” are established deserves at-
tention). In addition, there is a methodological-theoretical
working assumption, namely, that one can marry the study
of grammatical forms with conversation analysis. This work-
ing assumption has generated a substantial body of articles
and PhD dissertations, especially by students in applied lin-
guistics at UCLA and in linguistics at the University of Cal-
ifornia, Santa Barbara. Sidnell and Enﬁeld continue in this
tradition, this time tackling the old question of the impact
of linguistic structures on thinking and doing. They bring
some new data to the discussion of linguistic relativity and
present a hypothesis that can be further tested and debated.
For this they should be commended. Until recently, most of
the work on linguistic relativity has been either anecdotal or
experimental (Lucy 1992a, 1992b). The examination of nat-
urally occurring conversational interactions provides a much-
needed empirical testing ground for an important and yet
elusive issue. At the same time, in their choice of method,
analytical categories, and examples, the authors have also im-
plicitly and perhaps unintentionally supported a view of
social-action-as-language that ends up reducing the scope of
linguistic relativity and the potential impact of their work.
The problem, of course, lies not in the claim (made explicitly
at the beginning of the article) that words or utterances are
(meaningful) actions, something that is now generally ac-
cepted (even though sometimes forgotten) by linguists, phi-
losophers, and social scientists, but in the second part of the
article, where we are told that speciﬁc grammatical choices
have “nonlinguistic” effects and yet we ﬁnd out that those
effects are quite “linguistic”; for example, they are about
whether a topic is dropped or further expanded (through talk)
or about different perspectives (another language-mediated
This means that the authors have simultaneously widened
and narrowed the scope of the discussion of linguistic rela-
tivity. They have widened it by including an approach, con-
versation analysis, that offers ways of conceptualizing lan-
guage use that were not thinkable among the ﬁrst generations
of linguistic anthropologists, and they have narrowed it by
leaving out nonlinguistic actions or not engaging in an explicit
discussion of what they mean by “nonlinguistic.” Such a dis-
322 Current Anthropology Volume 53, Number 3, June 2012
cussion is necessary, in my view, because Whorf’s articlesshow
that he was comfortable speculating about ways in which
language patterns are connected to ways of doing things be-
yond language behavior, for example, ways of doing science,
ways of handling containers with ﬂammable substances, ways
of preparing for some future event.
Given the fact that both Sidnell and Enﬁeld have shown
themselves elsewhere to be interested in and sensitive to the
nonverbal aspects of social interaction, I was disappointed
that their “new program” of research did not include the idea
that “collateral effects” might include nonlinguistic actions or
(more likely) a combination of linguistic and nonlinguistic
actions, such as entering a house (from a particular side or
entrance), sitting down (in a particular spot), or looking at
others or avoiding eye-gaze—aspects of the interaction that
I know both authors are attuned to in their visual recordings
of interactions and that I also found important in determining
what happens next and how to interpret what is being said
or not said during otherwise quite routinized and predictable
interactions (e.g., Duranti 1992).
Finally, I want to mention another aspect of the relativity
issue that, in my view, needs attention in any new research
program, namely, the issue of freedom of action or, to use a
term more common in the social sciences, the issue of agency
(Duranti 2004, 2011). Sidnell and Enﬁeld’s discussion seems
to imply a certain inevitability of the collateral effects of lin-
guistic choices, but that does not have to be the case. In at
least one of the articles they cite (Heritage and Raymond,
forthcoming), it is shown that even though yes/no questions
restrain the scope of possible answers, speakers can resist the
“ﬁeld of constraints” established by the question and thus
assert their rights to go along with the request implicit or
explicit in the prior yes/no question.
Department of Sociology, University of California, Los Angeles,
264 Haines Hall, Los Angeles, California 90095-1551, U.S.A.
(email@example.com). 10 I 12
Universal Dilemmas and
In this fascinating paper, Sidnell and Enﬁeld take up a fairly
well understood Anglo-American interactional dilemma and
examine its management in other languages and cultures. The
dilemma arises from the fact that speakers of English and
other languages treat “going ﬁrst” as indexing greater episte-
mic authority and associate “going second” with lesser rights
and claims. The dilemma arising from this emerges when a
less epistemically entitled speaker evaluates some circum-
stance ahead of a more entitled one, thus setting the “terms
of agreement” for the more knowledgeable party and obliging
the latter to respond from a second, and incongruously sec-
ondary, position (Heritage and Raymond 2005). In the En-
glish case, resources for redressing this incongruity involve a
variety of practices that lower the epistemic claims of the ﬁrst
speaker and/or raise those of the second. In the following
case, Jenny, a family friend, evaluates Vera’s son Bill:
Jenny: Yeh .h well of course you see Bill is so good wih th’m ez
well is[n’t h[e:.
Vera: [.kl [That’s ri:ght yes.
In keeping with her inferior epistemic status with regard to
Bill, Jenny downgrades her evaluation with a tag question,
while Vera’s response asserts her superior epistemic rightswith
an explicit conﬁrmation positioned prior to a more acqui-
escent agreement token (“yes”; Raymond 2003). Here, the
“ofﬁcial” business of Vera’s turn is agreement, but the re-
sources through which the agreement is managed provide a
social lamination, coloring its meaning in terms of the epi-
stemic relations between the parties.
The English language has many expressions for this col-
oring of agreement: endorse, approve, concur, afﬁrm, con-
ﬁrm, and acquiesce, among others. However this epistemic
dilemma is probably not restricted to speakers of English.
Indeed, as Sidnell and Enﬁeld implicitly suggest, it may be a
quite general, possibly even a universal, one. In every culture,
persons who are “less expert” on the families, livestock, tools,
environments, histories, and experiences of others may ﬁnd
themselves “going ﬁrst” with an evaluation of those things,
thereby motivating their recipients to reassert their rights to
their own epistemic territory (Goffman 1971; Kamio 1997)
and to the social identities associated with those rights (Goff-
man 1959; Heritage 2011; Raymond and Heritage 2006; Sacks
1984). But if these kinds of epistemic incongruities are
widely—if not universally—experienced, they are nonetheless
indexed in an immense variety of ways. In some languages,
such as Japanese (Hayano 2011; Kamio 1997), speciﬁc par-
ticles are dedicated to indexing K⫹/K⫺epistemic relations,
while in others, such as English and, as Sidnell and Enﬁeld
illustrate, Finnish, Lao, and the Caribbean English Creole of
Bequia, a variety of more indirect means are deployed to the
The pragmatic logics of these indirect means areremarkably
various, even in the specimens to hand. While an Anglo-
American can assert epistemic priority by usurping the “ﬁrst-
ness” of a previous assessment through the production of a
second with a tag question or in negative interrogative format
(Heritage 2002b; Heritage and Raymond 2005), thereby oblig-
ing the ﬁrst speaker to respond in a newly created second
position, a resident of Bequia can achieve priority by re-
sponding to an assertion as if it were a question, thereby also
implicating a deﬁnitive ﬁnalization of the matter. Quite a
different avenue to this same outcome is achieved through
the Lao factive perfective le`q1.
As Durkheim (1995 ) taught, territories of knowl-
Sidnell and Enﬁeld Language Diversity and Social Action 323
edge and belief are intimately associated with the boundaries
of human groups. These boundaries and the actions that
make, sustain, reinforce, or defend them are sites where re-
lational indexing involves a conﬂuence of epistemics and af-
ﬁliation (Stivers, Mondada, and Steensig 2011). Some of the
linguistically collateral effects that Sidnell and Enﬁeld canvass
may emerge from the associations of speciﬁc practices with
afﬁliation and disafﬁliation. For example, English speakers
who wish to reply to questions while conveying thatthey were
redundant may preface the response with oh (Heritage 1998),
but similarly placed Mandarin speakers may sufﬁx a turn with
ﬁnal -a(Wu 2004). But ﬁnal -ais associated with oppositional
stances across a range of uses, and its use to index a redundant
question may, therefore, be more agonistic than its English
The Japanese (yo/ne) particle system effectively grammat-
icalizes the marking of relative epistemic authority, thus in-
stitutionalizing its relevance for speakers of that language.
Perhaps it is no accident, therefore, that it was the Japanese
scholar Akio Kamio who did so much to formulate the con-
cept of a territory of knowledge and to assert its relevance
for pragmatics. Perhaps patrolling epistemic territory is more
codable in Japanese or more cognitively salient for its speak-
ers. Perhaps pushing back on redundant questions is more
aggressive in Mandarin. Yet in Japanese, as in all languages,
there are multiple ways of marking epistemic territory, and
unless and until we have mapped out these ways and their
“collaterals” for each language, it will prove difﬁcult to reach
Sidnell and Enﬁeld’s enticing goal. Nonetheless, by focusing
their analysis on a sequentially delimited and socially similar
dilemma across sociocultural contexts, they have surely shown
us how the search can be prosecuted with some chance of
De´partement d’anthropologie, Universite´ de Montre´al, C.P. 6128,
Succursale Centre-Ville, Montre´ al, Que´bec H3C 3J7, Canada
(firstname.lastname@example.org). 2 II 12
This paper offers a welcome extension of the approaches hith-
erto taken to the implications of using a given language. The
key argument occurs toward the end of the text. “Because
[much] social action is done with the tools that our languages
provide and because these tools are structurally overdeter-
mined through their rich meanings and multiple functions,
the conventionalized selection of such tools will have lan-
guage-speciﬁc collateral effects on the ﬁnal nature of the ac-
tion.” This argument applies equally to other forms of lin-
guistic relativity and certainly to its classical form: to the
extent that we use language for conceptualization, we can do
so only in a particular language, and its distinctive forms will
shape—the shape analogy is as good as any—our thinking.
This kind of shaping is uncontroversial, for instance, for pho-
netics, giving what is called accent; the parallel is made clear
in John Lucy’s idea of semantic accent (e.g., Lucy 2003): the
semantic structure of a language “shaping” the meanings for
which it is used. This paper shows that not only sounds and
ideas but also social actions carry an accent.
What remains hard to sort out is what is a goal and what
is a collateral effect. For the authors, this does not seem to
be problematic, as they illustrate in a little parable: “First, one
has a certain end or goal: one wants to do something. Second,
in order to achieve that end or goal, one must select a means.”
This is perilously close to modern Western models of the
goal-oriented Homo oeconomicus, as well as to models of pure
thought existing before any symbolization (Leavitt 2011:50–
51). Do goals exist in some abstract form before all formu-
lation? Are they just things “one has”? Surely, our goals are
always already at least partly modeled by the means we have
used and expect to use: in Leonard Cohen’s words (from the
song “First We Take Manhattan”), we are “guided by the
beauty of our weapons.” In some cases, as in that of poetry
(Jakobson 1960), a central goal of the action is precisely the
maximization of the means. How much of social action is
more like the production of a poem, or what Le´vi-Strauss
(1962) called bricolage, than like a simple chain of given goal
and chosen tool?
The “expected” universality of some social goals (“target
action type[s] that we might expect speakers of any language
to want to carry out in social interaction”) is the basis for
the authors’ comparative framework; but what are these ex-
pectations based on? The example chosen, of agreeing with
a statement while asserting one’s greater authority to have
made the statement, would, for instance, simply not have been
part of the traditional linguaculture (Friedrich 1989)of north-
ern Athabaskan-speaking band societies. There, the appro-
priate response to an assertion made by someone with less
authority to make it than you have is silence. The kinds of
interactions found in the examples given here would have
represented an unacceptable violation of personal autonomy
(Guy Lanoue, personal communication).
This is not to deny the usefulness of pointing out a spe-
ciﬁcally social dimension of linguistic relativity: on the con-
trary. I would like to ﬁnish with an example. Many Indo-
Aryan languages feature the pervasive use of what are called
compound verbs: a concatenation of two verbs, the ﬁrst con-
veying the referential sense while the second, labeled the vec-
tor, light verb, or explicator verb, carries the grammatical
information and some kind of semantic or social nuance.
Among the most common such usages is that of the verbs
“give” and “take” to indicate the “benefactive” directionality
of the action toward someone else or toward oneself: “the
action of a transitive verb is potentially either for the beneﬁt
of the agent or for the ‘beneﬁt’ of someone or something
else—directed back toward or away from the agent, in other
words, but this remains latent in [New Indo-Aryan] until
speciﬁed by vectors take and give, even when . . . such in-
formation appears redundant” (Masica 1993:329). In Hindi,
324 Current Anthropology Volume 53, Number 3, June 2012
these are among the most commonly occurring vectors, and
their use seems to have been increasing, to the point that “de
‘give’ is coming to be the default means of marking an action
done for the benefaction of others; and le ‘take’ is coming to
mark benefaction for oneself” (Poornima and Painter 2010:
What “give” and “take” vectors add to an utterance is easy
to grasp in practice but hard to place linguistically: they mark
ﬂows of intentional action and allow (encourage?) a constant
explicit taking of responsibility for such ﬂows as a normal
part of discourse. How might such a linguistic resource inﬂect
the reputed intense interpersonal relationality of the societies
it marks (e.g., Roland 1988)? Sidnell and Enﬁeld’s third kind
of linguistic relativity gives us a theoretical frame for questions
of this kind.
Department of Linguistics; University of California, Berkeley, 1203
Dwinelle Hall, #2650, Berkeley, California 94720, U.S.A.
(email@example.com). 6 II 12
The Role of Systems in LR3
Sidnell and Enﬁeld’s delineation of a third locus for linguistic
relativity (LR3) constitutes a promising new front in our ef-
forts to understand the social signiﬁcance of language. Their
articulation of LR3 adumbrates a range of concrete empirical
questions that combines the study of linguistic form with that
of language use and of social action, thereby bridging the gulf
between linguistics and more socially oriented disciplines.
Sidnell and Enﬁeld’s articulation of LR3 clearly forms part
of the broader LR tradition, but their account of LR3 appears
to diverge from that tradition by downplaying a key aspect
of work focusing on the ﬁrst and second loci of linguistic
relativity (LR1 and LR2, respectively), namely, that LR effects
are due not only to the semantico-referential and/orindexical
properties of particular linguistic elements but also to the
participation of these elements in systems of elements. Sidnell
and Enﬁeld’s neglect of systems as such raises the question
of whether the kinds of communicative resources involved in
LR3 phenomena do not exhibit these systemic effects or
whether such effects remain signiﬁcant, even if they do not
loom large in the Sidnell and Enﬁeld’s discussion.
In the LR1 and LR2 traditions, the LR effects attributed
to, say, spatial-relational elements are explained not only in
terms of the meanings of speciﬁc elements but also in terms
of the properties of whole sets of semiotically related spatial-
reference terms. One of these properties is the “practical clo-
sure” that they exhibit with respect to particular communi-
cative tasks, namely, that they constitute, as a set, the principal
ready-at-hand means for achieving that task. Brieﬂy consider
a language like Guugu Yimidhirr, which employs a set of
elements to express spatial reference with respect to an ab-
solute frame of reference. Such a system exhibits practical
closure because although there may be means to express rel-
ative rather than absolute spatial reference, speakers do not
habitually avail themselves of those means. The second im-
portant property is the systemic property of the set of elements
in question. For example, in the case of Guugu Yimidhirr,
the speciﬁc lexicogrammatical elements used to express ab-
solute spatial reference stand in a variety of semiotic rela-
tionships with respect to each other, in classical structuralist
form: they are interlinked by sets of relations that constitute
the set of elements as a system that reﬂects a particular global
understanding of how space is organized (i.e., in terms of an
absolute frame of reference). Practical closure and these sys-
temic properties together guide users of such a system to
conceive of spatial relations in terms of absolute frames of
Although we may not necessarily expect the communicative
resources that Sidnell and Enﬁeld ﬁnd to be involved in LR3
phenomena to exhibit the same kind of systemic properties
as those involved in LR1 and LR2 phenomena (e.g., conver-
sation analysis does not usually emphasize “paradigmatic”
relationships among various interactional strategies), practical
closure and systemic properties presumably remain important
in LR3. Taking the speciﬁc case of K⫹second assessments
(K⫹2A), it is not clear that we can fully understand thesocial-
interactional effects of the linguistic resources employed in
expressing K⫹2A without an account of (1) the full set of
resources that speakers of a given language typically employ
to achieve K⫹2A and (2) the properties of that set of resources
as a system. Consider the use of if-prefaced repeats (IPRs) in
Guyanese Creole (GC) for K⫹2A, which is shown not only
to serve K⫹2A functions but also to exhibit topic-closing
functions. Without knowing the full set of K⫹2A resources,
it is unclear whether IPR is the only ready-at-hand resource
for K⫹2A and hence whether GC speakers are constrained
to simultaneously make a topic-closing move each time they
express K⫹2A. If the set of resources employed in GC K⫹2A
also includes resources that permit the expression of K⫹2A
without topic closing, the analysis of IPRs as having an un-
intended topic-closing effect is called into doubt. Crucially,
this issue cannot be resolved by examining IPRs alone but
requires examination of the complete set of communicative
resources that exhibit practical closure with respect to the
K⫹2A communicative function. On this view, LR3 effects
emerge not from speciﬁc communicative strategies, such as
IPR, but from the systemic properties of the set of functionally
related elements of which they form they form a part. It is
important to note that Sidnell’s and Enﬁeld’s discussion of
Lao and, especially, Finnish K⫹2A strategies appears to go
much farther toward an analysis of the full set of K⫹2A
strategies, but it is nevertheless interesting that the full extent
of K⫹2A resources is not clariﬁed in any of these cases. It
appears likely, however, that a comprehensive account of LR3
phenomena requires taking seriously the systemic properties
Sidnell and Enﬁeld Language Diversity and Social Action 325
of the set of elements that exhibit practical closure with respect
to the interactional move under consideration.
Department of Anthropology, University of Chicago, 1126 East
59th Street, Chicago, IL 60637-1580, U.S.A. (m-silverstein@
uchicago.edu). 5 II 12
Benjamin Whorf, a subtle master of the technical linguistics
of the 1930s, rethought the issues of “linguistic relativity” in
this light. He concerned himself with what we now term the
ontologization of linguistic forms, the “ontic commitments”
of speakers revealed in and by how they use certain “fashions
of speaking”—as we say, grammatically conforming collo-
cations. The question is whether—and how—speakers come
to project the categories evidenced by arrangements of form
in language into beliefs about the “reality” they experience
and conceptualize through linguistic denotation of them and
other semiotic behaviors.
Whorf went much farther than Boas, Sapir, and Bloomﬁeld
in this respect, even though he was limited by the focus of
the linguistics of his day on words and their internal structure
(or morphology). First, he extended the Boasian concept of
the “grammatical category”—a systematic formal coding of
some conceptual distinction (implying categorization of de-
notata)—to include not only those coded by ever-present,
transparent, and localizable pieces of the verbal signal. Whorf
pointed to subtle, implicit denotational categories immanent
in possible, but perhaps only occasionally occurring, conﬁg-
urations of the categories of the ﬁrst kind. His gradient in
two dimensions contrasted “overt, selective” categories and
“covert, modulus” categories. The reﬂexive consciousness of
well-intentioned laypersons focuses on the most overt and
selective categories in their ontologizations, in particular the
elements of the various major lexical classes that he termed
“lexations,” yielding such trivialities as the allegedly numerous
Inuit “words,” that is, simple noun roots, denoting phenom-
ena experienced as snow. (Boas himself hadlisted four.)Point-
ing out laypersons’ exceedingly limited reﬂexive consciousness
of covert, modulus categories, Whorf, like Boas, posited a
principled chasm between what speakers of a language actually
psychologically process—and inclusively code—about denot-
able “reality” and how they rely on their language’s relatively
overt and selective categories in their rationalizing ontological
claims about “what is ‘out there.’”
Second, in this light Whorf went on to posit a complex
dialectical process in which language plays a role in but does
not directly determine the enculturated intentionality of its
users, the “habitual thought world” that constitutes their de-
fault ontology—essentially “cultural concepts”—as they go
about their practically oriented social action. His argued this
by an extended example, comparing two such linguicultural
dialectics: (1) the “standard average European” (SAE) cultural
concept of “time” in our social universe of calendar and clock
reckoning, denoted by grammatical measure phrases, as
against the systematic semantics of SAE “tense” as a gram-
matical category, with which SAE natives confuse “time”; and
(2) the Hopi cultural concept of “emergent realizability” (tu-
natya) in the social universe of illocutionary forces of “making
manifest,” as against the systematic semantics of Hopi aspect
#modality #evidentiality as an intersected space of gram-
matical categories (Hopi lacks a grammatical category oftense
as such; see Whorf 1956 :213). For further detail, see
Silverstein (1979, 2000). The key idea here is the only partial
reﬂexive visibility of denotational category structures to
speakers and the consequent dialectical emergence of truly
culturally “relative” ontologizations of the “realities” that lan-
guage users denote.
Several of Whorf’s issues about language as an instrument
of denotation are also relevant when considering language as
a modality of “doing things with words,” that is, constituting
social action via its indexical (context-indicating) and iconic
(ﬁgurational) capacities, kinds of goal-oriented semiosis that
cross-cut but intersect its denotational functionality
though people may use language purposively to refer to en-
tities and to modally predicate states of affairs about them,
there is indeed a dialectic process that comes into view in
how people, using language seemingly to “talk about” mat-
ters—its generally recognizable denotational function
to coordination as sentient, agentive, identity-
bearing interlocutors through the effects of indexical signals
multifariously and pervasively patterned over the duration of
discourse. With vanishing rarity—except in courts and other
such institutions of state or church power—does one en-
counter a single turn at talk in which someone utters his or
her interactional purpose in speaking that very utterance (a
so-called explicit performative), and even then its effectiveness
as social action is hardly guaranteed (as even J. L. Austin
Sidnell and Enﬁeld’s transcripts in three different cultural
milieus all involve discursive acts of alignment to the deno-
tational content communicated in a prior turn by an inter-
acting other, although what we, as students of social inter-
action, can infer about what is going on in the way of actual
social action in each case is not treated in sufﬁcient ethno-
graphic detail to hazard more than a guess. Sidnell and Enﬁeld
contribute an interesting hypothesis in the realm of prag-
matics, parallel to Whorf’s demonstration (1956 :213)
in the realm of denotation that the Hopi category of “eye-
witness evidentiality” is an affordance for a speaker to denote
“pastness” by implying prior experience of whatever is re-
ported, a Gricean “implicature.” They suggest that the re-
spective forms in St. Vincent Creole, Finnish, and Lao by
which a speaker responds to an other’s earlier proposition,
seeming plus-or-minus to repeat it, indexes the current
speaker’s greater epistemic warrant or license to make such
a claim. Now it remains to investigate the hypothesis eth-
326 Current Anthropology Volume 53, Number 3, June 2012
Kathryn A. Woolard
Department of Anthropology, University of California, San Diego,
UCSD 0532, 9500 Gilman Drive, La Jolla, California 92039, U.S.A.
(firstname.lastname@example.org). 7 I 12
In wonderfully lucid prose, Sidnell and Enﬁeld give us a bril-
liant review of linguistic relativity and a welcome proposal
for a social approach to it. “Collateral effects” is a felicitous
term for the analogic processes and covert categories involved
in relativity. In admiration of the proposal’s systematicity and
scope and in sympathy with its social perspective, I have three
concerns: How new is it, how internally coherent, and how
well supported by evidence?
How different is Sidnell and Enﬁeld’s proposed form of rel-
ativity from Silversteinian indexicality; how different are the (if
I may) dependent variables “meaningful social behavior” and
“social action”? Agreeing while asserting authority sounds like
the kind of social action accomplishable with classic T/V al-
ternation. Sidnell and Enﬁeld hold that the difference lies in
their “speciﬁc sense of the sequences of interlocking actions
that people carry out,” while indexicality, in contrast, encom-
passes only a “general sense.” But does the different evidentiary
scale constitute a difference in kind of phenomenon? Although
indexicality studies rarely offer the microinteractional analysis
of conversation analysis (CA), the consequences of T/V selec-
tion can be shown unfolding similarly in interaction (Jacquemet
1994). We might even say it is the forms’ collateral effect to
appear to index general statuses.
Regarding the coherence of the argument, Sidnell and En-
ﬁeld’s principal claim morphs from strong proposals that
“different languages can have different effects on the kinds
of social actions that can be achieved” and “different lan-
guages can provide different opportunities for social action”
to the more modest proposal that although “the same” action
may be possible in different languages, the ways they differ
have implications for subsequent action. Sometimes the claim
seems tautological: the fact that different languages provide
different ways of doing an action has implications for how
that action is accomplished.
Which is demonstrated? To my eye, the extracts analyzed
here actually suggest that all three languages have a device for
“agreeing while asserting greater authority” that is similarly
closing-implicative; that is, the same social action and subse-
quent action are accomplished in all three, despite different
linguistic means. Sidnell and Enﬁeld emphasize that Finnish
has an alternative that allows topic elaboration instead of clo-
sure, but example (17) shows that Creole also has an alternative
(“I know ⫹repetition”) with a similar effect. Puzzlingly, this
is not incorporated into the analysis. It is not clear how this
comparison shows that there is a difference across languages
in the kind of social action that can be accomplished or even
a consequential difference for subsequent action.
Other puzzling bits of the argument can no doubt be
ﬂeshed out as it is developed more fully: collateral effects
derive from “other, nonequivalent functions,” but the word-
order patterns of Finnish are discussed only in their function
of agreeing with assessments, so it is difﬁcult to see how their
differential effects are collateral: to/from what? If in its “bap-
tismal” function (initiating repair) the Creole if-form is not
closing implicative, whence this collateral force, especially
when we know that there are three-part question events that
return the last word to the questioner (Mehan 1979)? And
why does a trace of a question form have such collateral force
in Creole, when in the Lao example (25), a full question is
dismissed as unhearable as anything other than rhetorical,
that is, as not having even the collateral, much less frontal,
force of its form?
Grammatical resources for social actions not only differ
across languages but also vary within languages (including
these), with more than one form for accomplishing a given
social action, as the authors acknowledge in footnotes. “Strat-
egies” for social actions, then, might be better located in
speakers than in languages. But if these are speakers’ strategic
choices (and do hearers not have choices, too?), then does
linguistic relativism once again recede as a will-o’-the-wisp?
Collateral effects depend on a failure of the semantic
bleaching that is generally seen in grammaticalization. With-
out assuming that origin is destiny or circularly treating sub-
sequent actions as both topic and resource, how do we know
that for native speakers the putative nuances of the baptismal
function of a form are brought along to its extended func-
tions? This leads to the general question of supporting evi-
dence. Since this is a programmatic piece, it is appropriate
that only a few illustrations are given. Rare as it is nowadays,
I would welcome at least the kind of summary characteri-
zation of the corpus that Schegloff (1968) gave of his “roughly
500” phone conversations, since in that classic CA work, one
deviant case led to a reanalysis. Although phrased in terms
of affordances, Sidnell and Enﬁeld’s thought-provoking pro-
posals about speciﬁc linguistic forms are still relatively causal
propositions that are prey to conﬁrmation bias and demand
testing against a body of data (not my ﬁeld’s strongest suit).
I am certain that these seasoned, innovative researchers can
provide such evidence, and I look forward to its exploration
in future work.
Department of Psychology and Centre for Situated Action and
Communication, University of Portsmouth, King Henry I Street,
Portsmouth PO1 2DY, United Kingdom (email@example.com).
Situated Action Is the Primary Locus of
Students of linguistics and psychology invariably ﬁnd the idea
that speakers of different languages “think differently” initially
fascinating. However, neo-Whorﬁan research (e.g., Boroditsky
Sidnell and Enﬁeld Language Diversity and Social Action 327
2001; Levinson 2003b; Lucy 1992a) has interpreted this in-
tuition in a very literal manner (with the book being cognitive
psychology), and in my experience, students’ fascination often
turns into bemusement or incredulity in the face of the pe-
culiar things psychologists ask their experimental participants
to do. I think that the approach outlined by Sidnell and
Enﬁeld, with its promise of a comparative science of human
social action, has the potential to point future students in-
terested in the idea of linguistic relativity in a direction that
will allow them to realistically explore the consequences of
There are many points of interest in this paper, and I want
to focus on the following. I want to ask, What will be the
main research question of the program of work sketched by
Sidnell and Enﬁeld? As researchers devote their energies to
studying the linguistic relativity of social action, this line of
work will likely become associated with one guiding question.
For example, most neo-Whorﬁan articles present their big
question as follows: Does language inﬂuence (affect, restruc-
ture) thought? What will the new question be for researchers
of grammar-related diversity in social action?
Sidnell and Enﬁeld formulate their question in different
ways. They ask, “Does the fact that the words were spoken
in English and not in another language have any bearing on
how that action [completing a howareyou sequence] is ac-
complished?” They also suggest, “[T]he language you speak
makes a difference for the social actions you can perform.”
These two ways of presenting the matter can be closelyrelated,
but they also have their distinct affordances. As an exercise
in exploring the linguistic relativity of social action, let us
consider the trajectories of thinking (and doing research) that
can be projected from these two formulations.
The ﬁrst question—Does grammar have a bearing on how
an action is accomplished?—can suggest that social actions
exist independently of the speciﬁc verbal and nonverbal acting
that brings them about and that local practices of speaking
might “affect” their realization. On this view, saying thatsocial
action is a “locus” of linguistic relativity can be taken to mean
that social action is one of the “dependent variables” onwhich
language exerts its “inﬂuence.” Along these lines, the research
program that Sidnell and Enﬁeld outline could join the neo-
Whorﬁan grooves of thinking, so that the question will be-
come Does language inﬂuence social action? I have argued
elsewhere that this way of asking the question (Does language
inﬂuence thought?) gets our thinking about language, culture,
and mind on the wrong track. It codiﬁes the abstraction and
decontextualization of “language,” “thought,” and “action”
from situated real-time speaking, thinking, and acting, and it
is accompanied by an entire glean of red herrings, such as
the classic supposed problem of the incommensurability of
The second question—Does grammar have a bearing on
what social actions speakers can perform?—can suggest that
talk in a particular grammatical format enters into the real-
time constitution of social actions. On this view, saying that
social action is a “locus” of linguistic relativity can be taken
to mean that it is in situated (social) activity that talk in a
particular grammatical format provides one of the material
resources for the moment-by-moment accomplishment of
cognition and action (Clark 2006; Goodwin 2000).
It is my impression that Sidnell and Enﬁeld have sympathies
for both of these ways in which research on the linguistic
relativity of social action might play out. Maybe it is possible
to keep both of these ways of thinking together. The line of
work discussed by Sidnell and Enﬁeld can open up a radically
different way of thinking about the relationship of linguistic
to cultural and cognitive diversity, if we manage to resist the
temptation of thinking of social action as merely a new de-
pendent variable and treat it instead as the primary site for
the situated development of cultural practice and cognitive
skills. On this view, situated social action is not so much a
third as the primary locus of linguistic relativity.
Collateral Effects, Agency, and Systems of
We sincerely thank the commentators for carefully engaging
with our work. We are encouraged by their support of this
line of research, and we are especially grateful for their critical
input toward clarifying our arguments.
We begin with “collateral effects.” While in the article we
focus on linguistic structures, both Duranti and Woolard
wonder about the breadth of scope of the idea. We expect to
see these effects in a broad range of domains and not only
the linguistic or “verbal.” Collateral effects are caused by de-
pendencies among the multiple features of any structure that
one may select for some function. Choosing a structure on
the basis of some subset of its features does not mean you
are exempt from “choosing” other features at the same time.
When you choose what car to buy, your choice may be guided
by speciﬁcations such as carrying capacity and fuel con-
sumption, but whether or not you care what color it is, you
cannot buy a car that is not some color. Suppose that while
you truly do not care about the color, the only car available
is canary yellow. A collateral effect of this purchase is that
the color, something you did not select but rather “settled
for,” will attract comments in a way that dark blue would not
have. Coming closer to language, Duranti asks whether we
see collateral effects in “nonverbal” behavior such as gestures.
Yes, we do. For instance, if you want to use your hand to
depict the path of motion of something you saw, you cannot
avoid showing the motion as occurring in a certain direction
(expressible as a cardinal direction such as north-northwest),
even when this is irrelevant to your expressive purpose (see
328 Current Anthropology Volume 53, Number 3, June 2012
Enﬁeld 2009a:17–18). Could such collateral effects of the
manual-visuospatial medium have consequences for social ac-
tion? As Duranti suggests, this is a matter for future research.
On Woolard’s query regarding the distinction between our
“third locus” of linguistic relativity and “Silversteinian in-
dexicality,” we refer the reader to the main article text, and
especially footnote 4, where we address this directly. Woolard
suggests that “agreeing while asserting authority sounds like
the kind of social action accomplishable with classic T/V al-
ternation.” Let us clarify why these are different. By “action”
here we mean “speech acts” such as requesting, inviting, of-
fering, complaining, excusing, agreeing, and disagreeing.
These can be done rudely or politely, with familiarity or dis-
tance, with either T or V pronoun forms. A T/V alternation
cannot accomplish an action in and of itself, although of
course some speciﬁc utterance containing a T oraVcan
accomplish the action of agreement (or K-plus agreement).
Unlike the indexical meanings associated with T and V forms,
the practices we discuss operate independently of the enduring
social relations of the parties and are in principle usable by
anyone, precisely because the K-plus versus K-minus distinc-
tion is always calibrated relative to some particular thing
In deﬁning collateral effects, we presuppose that people
have goals and that they select from among means to achieve
those goals. This strikes us as uncontroversial, although Lea-
vitt questions it. His worry cannot be that people do not have
goals. Think of the millions of big, middle-sized, and tiny
goals that ﬁll up your life. You need milk and there is none
in the fridge. Your goal: get some milk. To achieve it, you
could drive to the store, or you could call your friend who
is coming over and say, “Can you pick up some milk on the
way?” If Leavitt means that the matter is not always so simple,
then we agree. For one thing, goals can change as we go along.
For another, a single piece of behavior can fulﬁll multiple
goals. For yet another, culture does not just give us ways of
meeting our goals; it also speciﬁes the kinds of goals we
should, and sometimes must, try to meet. Choosing speciﬁc
solutions can introduce new, subordinate goals, and this is
another example of collateral effects. Leavitt also seems to
suggest that the determining of goals, and their selected so-
lutions, can be distributed across individuals and through
time in interaction. We agree that the matter of goals is nu-
anced in these ways, but this does not change our point.
People seldom act without reasons or without purpose, and
the means they select to meet their goals may introduce sec-
ondary, collateral effects.
If linguistic relativity effects are real, then different human
groups have different realms of possibility. Since our domain
of interest is action, this raises the issue of human “agency,”
as Duranti says. Our view is that collateral effects can be an
agency-reducing force, taking “agency” roughly to mean the
degree to which we can determine what we do and how we
do it (Duranti 1990; N. J. Enﬁeld, unpublished manuscript;
Kockelman 2007). Once you have chosen a strategy for certain
reasons, then your free will is in a sense now used up, and
you accept the collateral effects. PC or Mac? Rent or buy?
Chinese or Italian? Whatever the reason for your decision, a
higher-level choice will determine many other choices for you.
Woolard also points to agency in suggesting that strategies
for social actions “might be better located in speakers than
in languages.” But we are not forced to choose between people
and practices. Yes, there are people’s motives in speciﬁc cir-
cumstances, but then people have limited conventions to draw
on and hence not unlimited agency.
A question raised by both Heritage and Leavitt is whether
we expect the K⫹2A action to be a human universal. Our
reason for thinking that it probably is universal is that we
expect all human social groups to possess the basic ingredients
for this action: an economy of information, normative or-
ganization of rights and duties, and individuals’ motives to
maintain textured sets of social relations through practices of
(dis)afﬁliation. (For the ethnographic grounding of these as-
sumptions, see Sidnell 2005, chap. 2.) Cultures differ widely,
but there is a basic common infrastructure for social life that
is characteristic of our species (Enﬁeld and Levinson 2006a).
We can expect that in all cultures, individuals will be moti-
vated to form and maintain enduring personal relationships
of different types and that there will be information-related
practices such as agreeing with an evaluation someone else
just made (“They’re good kids,” “She can’t be trusted,” “This
is delicious”). Further, we can expect that in all cultures people
are motivated to “police” rights and duties (Henrich and Boyd
2001), in line with local norms, including those associated
with epistemic territory (Stivers, Mondada, and Steensig
2011). The nature and subtlety of such policing will vary, but
it will always occur, wherever norms are contravened (perhaps
most visibly in practices of socialization).
Managing epistemic rights and duties is surely done dif-
ferently in different cultures, and as Silverstein suggests, the
ethnographic context is important. But we do not share his
pessimism about the possibility of understanding the actions
exempliﬁed in our data. While Silverstein feels unsure about
“what is going on in the way of actual social action” in the
examples, there is no reason to think that the participants in
these interactions are any surer than he is. There is always a
rich ethnographic background, but what is its role in online
interpretation of social action? People’s understandings of
what is happening in the fast-moving enchronic context of
social life are never consummate and are seldom more than
adequate. Even in the coded meanings of words, people’s
understandings can differ without issue (Barr and Keysar
2005; Enﬁeld 2012b). Interaction is made possible by a
bounded form of rationality based on fast and frugal heuristics
(Gigerenzer, Hertwig, and Pachur 2011). There is no time for
elaborate ratiocination, especially as one needs to avoid the
implications associated with delayed response in interaction
(Stivers et al. 2009). So while ethnography is indispensible to
a full vision of social life, it cannot be that an encyclopedic
Sidnell and Enﬁeld Language Diversity and Social Action 329
knowledge of cultural context is invoked in an exhaustive
online interpretation of every bit of conduct in interaction.
A second issue concerning ethnography is the need to dis-
tinguish between ideologies regarding territories of knowl-
edge, which are sure to vary greatly across cultures, andactual,
nonreﬂective practice regarding territories of knowledge. In
discussing the proposed universality of the action type we are
examining, Leavitt offers a possible counterexample. He has
been told that in traditional “northern Athabaskan-speaking
band societies, . . . the appropriate response to an assertion
made by someone with less authority to make it than you
have is silence.” If silence is indeed the proper way to react
to a K-minus ﬁrst assertion in those societies, then it should
be recognizable as such. It would be readily describable in
the terms we suggested: while we showed, for example, that
a Creole solution to the K⫹2A problem is to treat the prior
turn as if it had been a question, this “silence” solution would
presumably treat the prior as if it had not been uttered at all.
We do not know whether this is what is happening in the
Athabaskan case, and indeed Leavitt implies that such se-
quences would not be found anyway, telling us that, as some-
one tells him, they “would have represented an unacceptable
violation of personal autonomy.” For now, we can only ex-
ercise the standard cautions in assessing this claim. Is it a
statement about what members of these societies say about
their linguistic practices, or is it about what they actually do?
At best, a known cultural ideology might lead to predictions
about what happens in interaction. To ﬁnd out, we would
need access to a corpus of recorded interaction in these so-
cieties, and we would begin by examining the corpus for
expressions of evaluation, especially the most mundane (e.g.,
“John’s new goats are pretty unruly”), and seeing how these
evaluations are expressed and taken up in relation to different
epistemic gradients (e.g., where “John” is related to the ad-
dressee, not the speaker).
Zinken is unhappy with our separation of action from lan-
guage, but such a separation is necessary for the simple reason
that social action is possible without language. Along lines
argued by Lucy (1992band elsewhere) for linguistic relativity
in relation to thought, it can be methodologically useful to
keep language and action apart so as to avoid tautology or
not to simply describe the same thing in two ways. Lee (1996
and elsewhere) argued against Lucy’s methodological sepa-
ration of language and thought (see also Hill and Mannheim
1992:382–385), proposing instead that everything is “langua-
thought.” Maybe this is what Zinken is going for as well:
everything is “langua-action.” But while language and action
are always connected in practice, so are fuel and vehicles,
guitars and music, or tools and carpentry. It does not mean
that you cannot distinguish the two conceptually in order to
study the phenomena. Social action is an important locus for
linguistic relativity, but it need not be the primary locus of
linguistic relativity (although perhaps there are arguments for
its primacy from speciﬁc perspectives, e.g., ontogenesis). The
phenomena we observe in social interaction would not take
place, and could not be understood, without all of these: social
action, language, culture, cognition, and more. Our concern
in writing the paper was to address a neglect of social action,
not to replace this with an equally problematic neglect of
some other equally necessary component of the story.
Finally, a recurring theme of the commentaries concerns
the systems in which sets of semiotic practices are embedded.
Duranti, Heritage, Michael, and Woolard all point out that
any language will supply not one but many possible forms to
choose from in formulating a social action and that individ-
uals’ agency should therefore be greater than we have implied.
While we agree that any language will provide a range of
options, central to our point is that the choices in one lan-
guage system are limited (see note 18). The scope of our article
was to describe the most prominent or common or idiomatic
ways of doing K-plus agreement in three language commu-
nities. For the Creole, for example, there are other ways (e.g.,
oh-prefacing), but these are less common and less central to
the vernacular (see note 19). Michael says that a full under-
standing of what we are claiming “requires examination of
the complete set of communicative resources that exhibit
practical closure with respect to the K⫹2A communicative
function.” We agree, and indeed, the situation is worse: we
would need to see all things that can happen after any ﬁrst-
This raises the issue of what a “grammar for interaction”
would look like, incorporating not only grammatical struc-
tures with their interactional uses but also interactional pat-
terns of preference and structural conformity and actions de-
ﬁned not in terms of morphosyntactic structures but in terms
of their distribution in conversational sequence (e.g., follow-
ing a K-minus initial assessment). Michael notes that con-
versation analysis “does not usually emphasize ‘paradigmatic’
relationships among various interactional strategies.” It is true
that this terminology has not been used, but in fact, the logic
of selecting from among members of paradigm sets is well
established in that literature. An example is the account of
person reference in terms of default versus marked forms
(Enﬁeld 2012a; Schegloff 1996b; Stivers, Enﬁeld, and Levinson
2007:9). Another is the analysis of the continuer uh huh,
which is in structural and functional opposition to two other
things that could happen in the same slot: initiating repair
and taking a speaking turn (Schegloff 1982; Sidnell 2009a).
Critically, the set of choices in any language is a subset of
the choices available in all the world’s languages together.
Whorf urged us to study languages so as to broaden our
worldviews. While nobody is “free to describe nature with
absolute impartiality,” he argued, the person who would come
closest “would be a linguist familiar with very many widely
different linguistic systems” (Whorf 1956 :214). This
is why we study human diversity: to know more about the
ways we can think and about the kinds of social systems we
can live in, and as we have argued here, to know more about
the human potential for action.
—N. J. Enﬁeld and Jack Sidnell
330 Current Anthropology Volume 53, Number 3, June 2012
Adelaar, Willem, and Pieter Muysken. 2007. The languages of the Andes. Cam-
bridge: Cambridge University Press.
Agha, Asif. 1994. Honoriﬁcation. Annual Review of Anthropology 23:277–302.
———. 2007. Language and social relations. Cambridge: Cambridge University
Aikhenvald, Alexandra. 2004. Evidentiality. Oxford: Oxford University Press
Atkinson, J. Maxwell, and John Heritage, eds. 1984. Structures of social action:
studies in conversation analysis. Cambridge: Cambridge University Press.
Austin, John L. 1962. How to do things with words. Oxford: Oxford University
Barr, Dale J., and Boaz Keysar. 2005. Making sense of how we make sense: the
paradox of egocentrism in language use. In Figurative language comprehension:
social and cultural inﬂuences. Herbert L. Colston and Albert N. Katz, eds. Pp.
21–41. Mahwah, NJ: Erlbaum.
Berlin, Brent, and Paul Kay.1969. Basic color terms: their universalityand evolution.
Berkeley: University of California Press.
Boas, Franz. 1889. On alternating sounds. American Anthropologist 2(1):47–51.
———. 1911. Introduction. In The handbook of American Indian languages.
Franz Boas, ed. Pp. 3–83. Washington, DC: Bureau of American Ethnology.
———. 1938. Language. In General anthropology. Franz Boas, ed. Pp. 124–145.
Boroditsky, Lera. 2001. Does language shape thought? English and Mandarin
speakers’ conceptions of time. Cognitive Psychology 43(1):1–22.
Boster, James S. 1985. Requiem for the omniscient informant: there’s life in the
old girl yet. In Directions in cognitive anthropology. Janet W. D. Dougherty,
ed. Pp. 177–197. Urbana: University of Illinois Press.
———. 1986. Exchange of varieties and information between Aguarunamanioc
cultivators. American Anthropologist 88(2):428–436.
Boster, James S., and Jeffrey C. Johnson. 1989. Form or function: a comparison
of expert and novice judgments of similarity among ﬁsh. American Anthro-
Boster, James S., Jeffrey C. Johnson, and Susan C. Weller. 1987. Social position
and shared knowledge: actors’ perceptions of status, role, and social structure.
Social Networks 9(4):375–387.
Briggs, Charles. 1986. Learning how to ask: a sociolinguistic appraisal of the role
of the interview in social science research. Cambridge: Cambridge University
Brown, Penelope, and Stephen C. Levinson. 1993. “Uphill” and “downhill” in
Tzeltal. Journal of Linguistic Anthropology 3(1):46–74.
Brown, Roger, and Albert Gilman. 1960. The pronouns of power and solidarity.
In Style in language. Thomas A. Sebeok, ed. Pp. 253–276. Cambridge, MA:
Clark, Andy. 2006. Language, embodiment, and the cognitive niche. Trends in
Cognitive Sciences 10(8):370–374. [JZ]
———. 2007. Supersizing the mind: embodiment, action, and cognitive extension.
Oxford: Oxford University Press.
Comrie, Bernard. 1981a.The languages of the Soviet Union. Cambridge: Cam-
bridge University Press.
———. 1981b.Language universals and linguistic typology. Oxford: Blackwell.
Curl, Traci S., and Paul Drew. 2008. Contingency and action: a comparison of
two forms of requesting. Research on Language and Social Interaction 41(2):
Dixon, R. M. W. 2002. Australian languages: their nature and development.Cam-
bridge: Cambridge University Press.
———. 2009. Basic linguistic theory, vols. 1 and 2. Oxford: Oxford University
Dixon, R. M. W., and Alexandra Y. Aikhenvald. 1999. The Amazonian languages.
Cambridge: Cambridge University Press.
Drew, Paul. 1991. Asymmetries of knowledge in conversational interactions. In
Asymmetries in dialogue. Ivana Markova` and Klaus Foppa, eds. Pp. 29–48.
Hemel Hampstead: Harvester Wheatsheaf.
———. 2004. Conversation analysis. In Handbook of language and social inter-
action. Kristine L. Fitch and Robert E. Sanders, eds. Pp. 71–102. Mahwah,
Drew, Paul, and Traci Walker. 2009. Going too far: complaining, escalating and
disafﬁliation. Journal of Pragmatics 41(12):2400–2414.
Duranti, Alessandro. 1990. Politics and grammar: agency in Samoan political
discourse. American Ethnologist 17(4):646–666.
———. 1992. Language and bodies in social space: Samoan ceremonial greetings.
American Anthropologist 94(3):657–691. [AD]
———. 1997. Linguistic anthropology. Cambridge: Cambridge University Press.
———. 1999. Relativity. Journal of Linguistic Anthropology 9(1–2):220–222.
———. 2003. Language as culture in U.S. anthropology. Current Anthropology
———. 2004. Agency in language. In A companion to linguistic anthropology.
Alessandro Duranti, ed. Pp. 451–473. Malden, MA: Blackwell. [AD]
———. 2009. Linguistic anthropology: history, ideas, and issues. In Linguistic
anthropology: a reader. Alessandro Duranti, ed. Pp. 1–60. Oxford: Blackwell.
———. 2011. Linguistic anthropology: language as a non-neutral medium. In
The Cambridge handbook of sociolinguistics. Rajend Mesthrie, ed. Pp. 28–46.
Cambridge: Cambridge University Press. [AD]
Durkheim, Emile. 1995 (1912). Elementary forms of the religious life. Karen Field,
trans. New York: Free Press. [JH]
Enﬁeld, N. J. 2005. Areal linguistics and mainland southeast Asia.Annual Review
of Anthropology 34:181–206.
———. 2006. Social consequences of common ground. In Roots of human
sociality: culture, cognition and interaction. N. J. Enﬁeld and Stephen C. Lev-
inson, eds. Pp. 399–430. Oxford: Berg.
———. 2007a.A grammar of Lao. Berlin: Mouton de Gruyter.
———. 2007b. Meanings of the unmarked: how “default” person referencedoes
more than just refer. In Person reference in interaction: linguistic, cultural and
social perspectives. N. J. Enﬁeld and Tanya Stivers, eds. Pp. 97–120. Cambridge:
Cambridge University Press.
———. 2009a.The anatomy of meaning: speech, gesture, and composite utterances.
Cambridge: Cambridge University Press.
———. 2009b. Relationship thinking and human pragmatics. Journal of Prag-
———. 2011a. Linguistic diversity in mainland Southeast Asia. In Dynamics of
human diversity: the case of mainland Southeast Asia. N. J. Enﬁeld, ed. Pp. 63–
80. Canberra: Paciﬁc Linguistics.
———. 2011b. Sources of asymmetry in human interaction: enchrony, status,
knowledge and agency. In The morality of knowledge in conversation. Tanya
Stivers, Lorenza Mondada, and Jakob Steensig, eds. Pp. 285–312. Cambridge:
Cambridge University Press.
———. 2012a. Reference in conversation. In Handbook of conversation analysis.
Jack Sidnell and Tanya Stivers, eds. Hoboken, NJ: Wiley.
———. 2012b. Tolerable friends. In Proceedings of the 33rd Annual Meeting of
the Berkeley Linguistics Society. Berkeley: University of California Press.
Enﬁeld, N. J., and Stephen C. Levinson. 2006a. Introduction: human sociality
as a new interdisciplinary ﬁeld. In Roots of human sociality: culture, cognition
and interaction. N. J. Enﬁeld and Stephen C. Levinson, eds. Pp. 1–35. Oxford:
———, eds. 2006b.Roots of human sociality: culture, cognition and interaction.
Errington, Joseph. 1985. On the nature of the sociolinguistic sign: describing
the Javanese speech levels. In Semiotic mediation: sociocultural and psychological
perspectives. Elizabeth Mertz and Richard J. Parmentier, eds. Pp. 287–310.
Orlando: Academic Press.
———. 1988. Structure and style in Javanese: a semiotic view of linguistic etiquette.
Philadelphia: University of Pennsylvania Press.
Evans, Nicholas, and Stephen C. Levinson. 2009. The myth of language uni-
versals: language diversity and its importance for cognitive science. Behavioral
and Brain Sciences 32(5):429–492.
Everett, Dan. 2005. Cultural constraints on grammar and cognition in Piraha˜:
another look at the design features of human language. Current Anthropology
Foley, William. 1986. The Papuan languages of New Guinea. Cambridge: Cam-
bridge University Press.
Friedrich, Paul. 1989. Language, ideology, and political economy. American An-
thropologist 91(2):295–312. [JL]
Gentner, Dedre. 1982. Why nouns are learned before verbs: linguistic relativity
versus natural partitioning. In Language, thought and culture, vol. 2 of Language
development. Stan A. Kuczaj, ed. Pp. 301–334. Hillsdale, NJ: Erlbaum.
Gentner, Dedre, and Susan Goldin-Meadow. 2003. Whither Whorf. InLanguage
in mind: advances in the study of language and thought. Dedre Gentner and
Susan Goldin-Meadow, eds. Pp. 3–14. Cambridge, MA: MIT Press.
Gentner, Dedre, Mutsumi Imai, and Lera Boroditsky. 2002. As time goes by:
evidence for two systems in processing spacertime metaphors. Language and
Cognitive Processes 17(5):537–565.
Gigerenzer, Gerd, Ralph Hertwig, and Thorsten Pachur, eds. 2011. Heuristics:
the foundations of adaptive behavior. New York: Oxford University Press.
Goffman, Erving. 1959. The presentation of self in everyday life. Garden City, NY:
Sidnell and Enﬁeld Language Diversity and Social Action 331
———. 1971. Relations in public: microstudies of the public order.NewYork:
Harper & Row. [JH]
———. 1981. Footing. In Forms of talk. Pp. 124–159. Philadelphia: University
of Pennsylvania Press.
Goodwin, Charles. 2000. Action and embodiment within situated human in-
teraction. Journal of Pragmatics 32(10):1489–1522.
———. 2002. Time in action. Current Anthropology 43(suppl.):S19–S35.
———. 2003. Conversational frameworks for the accomplishment of meaning
in aphasia. In Conversation and brain damage. Charles Goodwin, ed. Pp. 90–
116. Oxford: Oxford University Press.
———. 2006. Human sociality as mutual orientation in a rich interactive en-
vironment: multimodal utterances and pointing in aphasia. In Roots of human
sociality: culture, cognition and interaction. N. J. Enﬁeld and Stephen C. Lev-
inson, eds. Pp. 96–125. Oxford: Berg.
Goodwin, Charles, and Marjorie Harness Goodwin. 1987. Concurrentoperations
on talk: notes on the interactive organization of assessments. IPrA Papers in
Goodwin, Charles, and John Heritage. 1990. Conversation analysis. Annual Re-
view of Anthropology 19:283–307.
Gumperz, John J., and Stephen C. Levinson, eds. 1996. Rethinking linguistic
relativity. Cambridge: Cambridge University Press.
Hakulinen, Auli, and Marja-Leena Sorjonen. 2009. Designing utterances for
action: ways of repeating the verb as a response to an assessment. In Talk-in-
interaction: comparative dimensions. Markku Haakana, Minna Laakso, and Jan
Lindstro¨m, eds. Pp. 124–151. Helsinki: Finnish Literature Society.
Hanks, William F. 1990. Referential practice: language and lived space among the
Maya. Chicago: University of Chicago Press.
———. 1992. The indexical ground of deictic reference. In Rethinking context:
language as an interactive phenomenon. Alessandro Duranti and Charles Good-
win, eds. Pp. 43–77. Cambridge: Cambridge University Press.
———. 1993. Metalanguage and pragmatics of deixis. In Reﬂexive language:
reported speech and metapragmatics. John A. Lucy, ed. Pp. 127–157. Cambridge:
Cambridge University Press.
———. 2005a. Explorations in the deictic ﬁeld. Current Anthropology 46(2):
———. 2005b. Pierre Bourdieu and the practices of language. Annual Review
of Anthropology 34:67–83.
———. 2006. Spatial frames of reference in language and thought. Language in
Haspelmath, Martin, Matthew S. Dryer, David Gil, and Bernard Comrie, eds.
2005. The world atlas of language structures. Oxford: Oxford University Press.
Haviland, John B. 1979. Guugu Yimidhirr brother-in-law language. Language in
———. 1993. Anchoring, iconicity, and orientation in Guugu Yimithirr pointing
gestures. Journal of Linguistic Anthropology 3(1):3–45.
———. 1996. Projections, transpositions, and relativity. In Rethinking linguistic
relativity. John J. Gumperz and Stephen C. Levinson, eds. Pp. 271–323. Cam-
bridge: Cambridge University Press.
———. 1998. Guugu Yimithirr cardinal directions. Ethos 26(1):25–47.
———. 2000. Pointing, gesture spaces, and mental maps. In Language and
gesture: window into thought and action. David McNeill, ed. Pp. 13–46. Cam-
bridge: Cambridge University Press.
Hayano, Kaoru. 2011. Claiming epistemic primacy: yo-marked assessments in
Japanese. In The morality of knowledge in conversation. Tanya Stivers, Lorenza
Mondada, and Jakob Steensig, eds. Pp. 58–81. Cambridge: Cambridge Uni-
versity Press. [JH]
———. 2012. Territories of knowledge in Japanese conversation. PhD disser-
tation, Max Planck Institute for Psycholinguistics, Nijmegen.
Henrich, Joseph, and Robert Boyd. 2001. Why people punish defectors: weak
conformist transmission can stabilize costly enforcement of norms in coop-
erative dilemmas. Journal of Theoretical Biology 208(1):79–89.
Herder, Johann Gottfried. 2002 (1772). Treatise on the origin of language. In
Herder: philosophical writings. Michael N. Forster, trans. and ed. Pp. 65–164.
Cambridge: Cambridge University Press.
Heritage, John. 1984. Garﬁnkel and ethnomethodology. Cambridge: Polity Press.
———. 1998. Oh-prefaced responses to inquiry. Language in Society 27(3):291–
———. 2002a.Oh-prefaced responses to assessments: a method of modifying
agreement/disagreement. In The language of turn and sequence. Cecilia E. Ford,
Barbara A. Fox, and Sandra A. Thompson, eds. Pp. 196–224. Oxford and
New York: Oxford University Press.
———. 2002b. The limits of questioning: negative interrogatives and hostile
question content. Journal of Pragmatics 34(10–11):1427–1446. [JH]
———. 2011. Territories of knowledge, territories of experience: empathic mo-
ments in interaction. In The morality of knowledge in conversation. Tanya
Stivers, Lorenza Mondada, and Jakob Steensig, eds. Pp. 159–183. Cambridge:
Cambridge University Press. [JH]
Heritage, John, and Geoffrey Raymond. 2005. The terms of agreement: indexing
epistemic authority and subordination in talk-in-interaction.Social Psychology
———. Forthcoming. Navigating epistemic landscapes: acquiescence, agency
and resistance in responses to polar questions. In Questions: formal, functional
and interactional perspectives. Jan P. de Ruiter, ed. Cambridge: Cambridge
Hill, Jane H., and Bruce Mannheim. 1992. Language and world-view. Annual
Review of Anthropology 21:381–406.
Humboldt, Wilhelm von. 1999 (1836). On language: the diversity of human
language construction and its inﬂuence on the mental development of thehuman
species. Peter Heath, trans. Michael Losonsky, ed. Cambridge: Cambridge
Hunt, Earl, and Franca Agnoli. 1991. The Whorﬁan hypothesis: a cognitive
psychology perspective. Psychological Review 98(3):377–389.
Hutchins, Edwin. 1995. Cognition in the wild. Cambridge, MA: MIT Press.
———. 2006. The distributed cognition perspective on human interaction. In
Roots of human sociality: culture, cognition and interaction.N.J.Enﬁeldand
Stephen C. Levinson, eds. Pp. 375–398. Oxford: Berg.
Hymes, Dell. 1966. Two types of linguistic relativity (with examples from Am-
erindian ethnography). In Sociolinguistics: proceedings of the UCLA Sociolin-
guistics Conference, 1964. William Bright, ed. Pp. 114–157. The Hague: Mou-
Imai, Mutsumi, and Dedre Gentner. 1993. Linguistic relativity vs. universal
ontology: cross-linguistic studies of the object/substance distinction. In Pa-
pers from the 29th Regional Meeting of the Chicago Linguistics Society, vol.
2. What we think, what we mean and how we say it: papers from the pa-
rasession on the correspondence of conceptual, semantic and grammatical rep-
resentations. Katharine Beals, Gina Cooke, David Kathman, Karl-Erik
McCullough, Sotaro Kita, and David Testen, eds. Pp. 171–186. Chicago:
Chicago Linguistics Society.
Jacquemet, Marco. 1994. T-offenses and metapragmatic attacks: strategies of
interactional dominance. Discourse and Society 5(3):297–319.
Jakobson, Roman. 1960. Linguistics and poetics. In Style in language. Thomas
A. Sebeok, ed. Pp. 350–377. Cambridge, MA: MIT Press. [JL]
———. 1971 (1957). Shifters, verbal categories, and the Russian verb. InSelected
writings of Roman Jakobson. Vol. 2. Pp. 130–147. The Hague: Mouton.
———. 1971 (1959). Boas’ view of grammatical meaning. In Selected writings
of Roman Jakobson. Vol. 2. Pp. 489–496. The Hague: Mouton.
Jefferson, Gail. 1980. On “trouble-premonitory” response to inquiry. Sociological
Kamio, Akio. 1997. Territory of information. Amsterdam: Benjamins. [JH]
Kay, Paul, and Willett Kempton. 1984. What is the Sapir-Whorf hypothesis?
American Anthropologist 86(1):65–79.
Kockelman, Paul. 2005. The semiotic stance. Semiotica 157(1/4):233–304.
———. 2007. Agency: the relation between meaning, power, and knowledge.
Current Anthropology 48(3):375–401.
Koerner, E. F. Konrad. 1992. The Sapir-Whorf hypothesis: a preliminary history
and a bibliographic essay. Journal of Linguistic Anthropology 2(2):173–178.
Krishnamurti, Bhadriraju. 2003. The Dravidian languages. Cambridge: Cam-
bridge University Press.
Lave, Jean. 1988. Cognition in practice: mind, mathematics and culture in everyday
life. Cambridge: Cambridge University Press.
Lave, Jean, and Etienne Wenger. 1991. Situated learning: legitimate peripheral
participation. Cambridge: Cambridge University Press.
Leavitt, John. 2006. Linguistic relativities. In Language, culture and society. Chris-
tine Jourdan and Kevin Tuite, eds. Pp. 47–81. Cambridge: Cambridge Uni-
———. 2011. Linguistic relativities: language diversity and modern thought. Cam-
bridge: Cambridge University Press.
Lee, Benjamin. 1985. Peirce, Frege, Saussure, and Whorf: the semiotic mediation
of ontology. In Semiotic mediation: sociocultural and psychological perspectives.
Elizabeth Mertz and Richard J. Parmentier, eds. Pp. 99–128. Orlando, FL:
Lee, Penny. 1991. Whorf’s Hopi tensors: subtle articulators in the language/
thought nexus? Cognitive Linguistics 2(2):123–147.
———. 1996. The Whorf theory complex: a critical reconstruction. Amsterdam:
———. 2000. When is linguistic relativity Whorf’s linguistic relativity? In Ex-
332 Current Anthropology Volume 53, Number 3, June 2012
plorations in linguistic relativity. Martin Pu¨ tz and Marjolijn J. Verspoor, eds.
Levinson, Stephen C. 1983. Pragmatics. Cambridge: Cambridge University Press.
———. 1992. Primer for the ﬁeld investigation of spatial description and con-
ception. Pragmatics 2(1):5–47.
———. 1996a. Frames of reference and Molyneux’s question: cross-linguistic
evidence. In Language and space. Paul Bloom, Mary A. Peterson, Lynn Nadel,
and Merrill F. Garrett, eds. Pp. 109–169. Cambridge, MA: MIT Press.
———. 1996b. Language and space. Annual Review of Anthropology 25:353–382.
———. 1996c. Relativity in spatial conception and description. In Rethinking
linguistic relativity. John J. Gumperz and Stephen C. Levinson, eds. Pp. 177–
202. Cambridge: Cambridge University Press.
———. 1997a. From outer to inner space: linguistic categories and non-linguistic
thinking. In Language and conceptualization. Jan Nuyts and Eric Pederson,
eds. Pp. 13–45. Cambridge: Cambridge University Press.
———. 1997b. Language and cognition: the cognitive consequences of spatial
description in Guugu Yimithirr. Journal of Linguistic Anthropology 7(1):1–35.
———. 2000. Ye´lıˆ Dnye and the theory of basic color terms. Journal of Linguistic
———. 2001. Covariation between spatial language and cognition. In Language
acquisition and conceptual development. Melissa Bowerman and Stephen C.
Levinson, eds. Pp. 566–588. Cambridge: Cambridge University Press.
———. 2003a. Language and mind: let’s get the issues straight! In Language in
mind: advances in the study of language and cognition. Dedre Gentner and
Susan Goldin-Meadow, eds. Pp. 25–46. Cambridge, MA: MIT Press.
———. 2003b.Space in language and cognition. Cambridge: Cambridge Uni-
———. 2005. Living with Manny’s dangerous idea. Discourse Studies 7(4–5):
Levinson, Stephen C., and Nicholas Evans. 2010. Time for a sea-change in
linguistics: response to comments on “The Myth of Language Universals.”
Le´ vi-Strauss, Claude. 1962. La pense´e sauvage. Paris: Plon. [JL]
Lucy, John A. 1985. Whorf’s view of the linguistic mediation of thought. In
Semiotic mediation: sociocultural and psychological perspectives. Elizabeth Mertz
and Richard J. Parmentier, eds. Pp. 73–97. Orlando, FL: Academic Press.
———. 1992a.Grammatical categories and cognition: a case study of the linguistic
relativity hypothesis. Cambridge: Cambridge University Press.
———. 1992b.Language diversity and thought: a reformulation of the linguistic
relativity hypothesis. Cambridge: Cambridge University Press.
———. 1996. The scope of linguistic relativity: an analysis and review of em-
pirical research. In Rethinking linguistic relativity. John J. Gumperz and Stephen
C. Levinson, eds. Pp. 37–69. Cambridge: Cambridge University Press.
———. 1997. Linguistic relativity. Annual Review of Anthropology 26:291–312.
———. 2003. Semantic accent and linguistic relativity. Paper presented at the
Conference on Cross-Linguistic Data and Theories of Meaning. Max Planck
Institute for Psycholinguistics, Nijmegen, Netherlands, May 20. [JL]
Lucy, John A., and Richard A. Shweder. 1979. Whorf and his critics: linguistic
and nonlinguistic inﬂuences on color memory. American Anthropologist 81(3):
Luong, Hy Van. 1984. “Brother” and “uncle”: an analysis of rules, structural
contradictions, and meaning in Vietnamese kinship. American Anthropologist
———. 1987. Plural markers and personal pronouns in Vietnamese person
reference: an analysis of pragmatic ambiguity and native. Anthropological Lin-
———. 1988. Discursive practices and power structure: person-referring forms
and sociopolitical struggles in colonial Vietnam. American Ethnologist 15(2):
———. 1990. Discursive practices and linguistic meanings: the Vietnamese system
of person reference. Amsterdam: Benjamins.
Majid, Asifa, James S. Boster, and Melissa Bowerman. 2008. The cross-
linguistic categorization of everyday events: a study of cutting and breaking.
Majid, Asifa, Melissa Bowerman, Sotaro Kita, Daniel B. M. Haun, and Stephen
C. Levinson. 2004. Can language restructure cognition? the case for space.
Trends in Cognitive Sciences 8(3):108–114.
Masica, Colin P. 1993. The Indo-Aryan languages. Cambridge: Cambridge Uni-
versity Press. [JL]
Mehan, Hugh. 1979. “What time is it, Denise?” asking known information
questions in classroom discourse. Theory into Practice 28(4):285–294. [KAW]
Mithun, Marianne. 2001. The languages of native North America. Cambridge:
Cambridge University Press.
Moerman, Michael, and Harvey Sacks. 1988 (1970). On “understanding” in the
analysis of natural conversation. In Talking culture: ethnography and conver-
sation analysis. Pp. 180–186. Philadelphia: University of Pennsylvania Press.
Ochs, Elinor. 1988. Culture and language development: language acquisition and
language socialization in a Samoan village. Cambridge: Cambridge University
———. 1990. Indexicality and socialization. In Cultural psychology: essays on
comparative human development. James W. Stigler, Richard A. Shweder, and
Gilbert Herdt, eds. Pp. 287–308. Cambridge: Cambridge University Press.
———. 1992. Indexing gender. In Rethinking context: language as an interactive
phenomenon. Alessandro Duranti and Charles Goodwin, eds. Pp. 335–358.
Cambridge: Cambridge University Press.
———. 1996. Linguistic resources for socializing humanity. In Rethinking lin-
guistic relativity. John J. Gumperz and Stephen C. Levinson, eds. Pp. 407–
437. Cambridge: Cambridge University Press.
Parmentier, Richard J. 1994a. Peirce divested for non-initiates. InSigns in society:
studies in semiotic anthropology. Pp. 3–22. Bloomington: Indiana University
———. 1994b. Peirce’s concept of semiotic mediation. In Signs in society: studies
in semiotic anthropology. Pp. 23–44. Bloomington: Indiana University Press.
Pederson, Eric, Eve Danziger, David Wilkins, Stephen Levinson, SotaroKita, and
Gunter Senft. 1998. Semantic typology and spatial conceptualization.Language
Peirce, Charles Sanders. 1955. Logic as semiotic: the theory of signs. In Philo-
sophical writings of Peirce. Justus Buchler, ed. Pp. 98–120. New York: Dover.
Pomerantz, Anita. 1978. Compliment responses: notes on the co-operation of
multiple constraints. In Studies in the organization of conversational interaction.
Jim Schenkein, ed. Pp. 79–112. New York: Academic Press.
———. 1984. Agreeing and disagreeing with assessments: some features of
preferred/dispreferred turn shapes. In Structures of social action: studies in
conversation analysis. J. Maxwell Atkinson and John Heritage, eds. Pp. 57–
101. Cambridge: Cambridge University Press.
Poornima, Shakthi, and Robert Painter. 2010. Divergent pathways: the current
status of grammaticalization of Hindi verbs. In Yearbook of South Asian lan-
guages and linguistics. Rajendra Singh, ed. Pp. 75–106. Berlin: de Gruyter. [JL]
Posner, Rebecca. 1996. The Romance languages. Cambridge: Cambridge Uni-
Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech, and Jan Svartvik. 1985.
A comprehensive grammar of the English language. New York: Longman.
Raymond, Geoffrey. 2003. Grammar and social organization: yes/no interrog-
atives and the structure of responding. American Sociological Review 68(6):
Raymond, Geoffrey, and John Heritage. 2006. The epistemics of social relations:
owning grandchildren. Language in Society 35(5):677–705.
Roland, Alan. 1988. In search of self in India and Japan: toward a cross-cultural
psychology. Princeton, NJ: Princeton University Press. [JL]
Romney, A. Kimball. 1999. Culture consensus as a statistical model. Cultural
Romney, A. Kimball, William H. Batchelder, and Susan C. Weller. 1987. Recent
applications of consensus theory. American Behavioral Scientist 31(2):163–177.
Romney, A. Kimball, Susan C. Weller, and William H. Batchelder. 1986. Culture
as consensus: a theory of culture and informant accuracy. American Anthro-
Rumsey, Alan. 1990. Wording, meaning, and linguistic ideology. American An-
Sacks, Harvey. 1975. Everyone has to lie. In Sociocultural dimensions of language
use. Mary Sanches and Ben G. Blount, eds. Pp. 57–80. New York: Academic
———. 1984. On doing “being ordinary.” In Structures of social action.J.Maxwell
Atkinson and John Heritage, eds. Pp. 413–429. Cambridge: Cambridge Uni-
versity Press. [JH]
———. 1987. On the preferences for agreement and contiguity in sequences in
conversation. In Talk and social organisation. Graham Button and John R. E.
Lee, eds. Pp. 54–69. Clevedon, England: Multilingual Matters.
———. 1995. Lectures on conversation, 2 vols. Oxford: Blackwell.
Sacks, Harvey, Emanuel A. Schegloff, and Gail Jefferson. 1974. A simplest sys-
tematics for the organization of turn-taking for conversation. Language 50(4):
Sahlins, Marshall. 1976. Culture and practical reason. Chicago: University of
Sapir, Edward. 1921. Language: an introduction to the study of speech.NewYork:
———. 1949 (1927). The unconscious patterning of behavior in society. In
Sidnell and Enﬁeld Language Diversity and Social Action 333
Selected writings of Edward Sapir. David G. Mandelbaum, ed. Pp. 544–555.
Berkeley: University of California Press.
———. 1964 (1931). Conceptual categories in primitive languages. InLanguage
in culture and society: a reader in linguistics and anthropology. Dell Hymes, ed.
Pp. 128. New York: Harper & Row.
Schegloff, Emanuel A. 1968. Sequencing in conversational openings. American
———. 1982. Discourse as an interactional achievement: some uses of“uh huh”
and other things that come between sentences. In Analyzing discourse: text
and talk. Deborah Tannen, ed. Pp. 71–93. GeorgetownUniversity Round Table
on Languages and Linguistics 1981. Washington, DC: Georgetown University
———. 1986. The routine as achievement. Human Studies 9(2–3):111–151.
———. 1991. Conversation analysis and socially shared cognition. In Perspectives
on socially shared cognition. Lauren B. Resnick, John M. Levine, and Stephanie
D. Teasley, eds. Pp. 150–171. Washington DC: American Psychological As-
———. 1992. Repair after next turn: the last structurally provided for place for
the defense of intersubjectivity in conversation. American Journal of Sociology
———. 1996a. Conﬁrming allusions: toward an empirical account of action.
American Journal of Sociology 104(1):161–216.
———. 1996b. Some practices for referring to persons in talk-in-interaction: a
partial sketch of a systematics. In Studies in anaphora. Barbara Fox, ed. Pp.
437–485. Amsterdam: Benjamins.
———. 2006. Interaction: the infrastructure for social institutions, the natural
ecological niche for language, and the arena in which culture is enacted. In
Roots of human sociality: culture, cognition and interaction.N.J.Enﬁeldand
Stephen C. Levinson, eds. Pp. 70–96. Oxford: Berg.
———. 2007. Sequence organization in interaction: a primer in conversation anal-
ysis, vol. 1. Cambridge: Cambridge University Press.
Schegloff, Emanuel A., Gail Jefferson, and Harvey Sacks. 1977. The preference
for self-correction in the organization of repair in conversation. Language
Shibatani, Masayoshi. 1990. The languages of Japan. Cambridge: Cambridge Uni-
Shopen, Timothy, ed. 1985. Language typology and syntactic description, 3 vols.
Cambridge: Cambridge University Press.
———, ed. 2007. Language typology and syntactic description, 3 vols. 2nd edition.
Cambridge: Cambridge University Press.
Sidnell, Jack. 2005. Talk and practical epistemology: the social life of knowledge in
a Caribbean community. Amsterdam: Benjamins.
———. 2007a. Comparative studies in conversation analysis. Annual Review of
———. 2007b. Repairing person reference in a small Caribbean community. In
Person reference in interaction: linguistic, cultural and social perspectives.N.J.
Enﬁeld and TanyaStivers, eds. Pp. 281–308. Cambridge: Cambridge University
———. 2009a. Comparative perspectives in conversation analysis. In Conver-
sation analysis: comparative perspectives. Jack Sidnell, ed. Pp. 3–33. Cambridge:
Cambridge University Press.
———, ed. 2009b.Conversation analysis: comparative perspectives. Cambridge:
Cambridge University Press.
———. 2009c. Language-speciﬁc resources in repair and assessments. In Con-
versation analysis: comparative perspectives. Jack Sidnell, ed. Pp. 303–320. Cam-
bridge: Cambridge University Press.
———. 2010. Conversation analysis: an introduction. Oxford: Blackwell/Wiley.
Silverstein, Michael. 1976. Shifters, linguistic categories, and culturaldescription.
In Meaning in anthropology. Keith H. Basso and Henry A. Selby, eds. Pp. 11–
55. Albuquerque: University of New Mexico Press.
———. 1979. Language structure and linguistic ideology. In The elements: a
parasession on linguistic units and levels. Paul R. Clyne, William F. Hanks, and
Carol L. Hofbauer, eds. Pp. 193–247. Chicago: Chicago Linguistic Society.
———. 1981. The limits of awareness. Working Papers in Sociolinguistics, no.
84. Austin, TX: Southwest Educational Laboratory.
———. 1985. Language and the culture of gender: at the intersection of struc-
ture, usage, and ideology. In Semiotic mediation: sociocultural and psychological
perspectives. Elizabeth Mertz and Richard J. Parmentier, eds. Pp. 219–259.
Orlando, FL: Academic Press.
———. 1987. The three faces of “function”: preliminaries to a psychology of
language. Social and functional approaches to language and thought. Maya
Hickmann, ed. Pp. 17–38. Orlando, FL: Academic Press.
———. 2000. Whorﬁanism and the linguistic imagination of nationality. In
Regimes of language: ideologies, polities, identities. Paul V. Kroskrity, ed. Pp.
85–138. Santa Fe, NM: School of American Research Press. [MS]
———. 2003. Indexical order and the dialectics of sociolinguistic life. Language
and Communication 23(3–4):193–229.
———. 2004. “Cultural” concepts and the language-culture nexus. Current
Slobin, Dan I. 1996. From “thought and language” to “thinking for speaking.”
In Rethinking linguistic relativity. John J. Gumperz and Stephen C. Levinson,
eds. Pp. 70–96. Cambridge: Cambridge University Press.
Sorjonen, Marja-Leena. 1996. On repeats and responses in Finnish conversations.
In Interaction and grammar. Elinor Ochs, Emanuel A. Schegloff, and Sandra
A. Thompson, eds. Pp. 277–327. Cambridge: Cambridge University Press.
Sorjonen, Marja-Leena, and Auli Hakulinen. 2009. Alternative responses to as-
sessments. In Conversation analysis: comparative perspectives. Jack Sidnell, ed.
Pp. 281–303. Cambridge: Cambridge University Press.
Stivers, Tanya. 2005. Modiﬁed repeats: one method for asserting primary rights
from second position. Research on Language and Social Interaction 38(2):131–
———. 2011. Morality and question design: “of course” as contesting a pre-
supposition of askability. In The morality of knowledge in conversation. Tanya
Stivers, Lorenza Mondada, and Jakob Steensig, eds. Cambridge: Cambridge
Stivers, Tanya, N. J.Enﬁeld, Penelope Brown, Christina Englert, Makoto Hayashi,
Trine Heinemann, Gertie Hoymann, et al. 2009. Universals and cultural var-
iation in turn-taking in conversation. Proceedings of the National Academy of
Sciences of the USA 106(26):10587–10592.
Stivers, Tanya, N. J. Enﬁeld, and Stephen C. Levinson. 2007. Person reference
in interaction. In Person reference in interaction: linguistic, cultural and social
perspectives. N. J. Enﬁeld and Tanya Stivers, eds. Pp. 1–20. Cambridge: Cam-
bridge University Press.
Stivers, Tanya, Lorenza Mondada, and Jakob Steensig, eds. 2011. The morality
of knowledge in conversation. Cambridge: Cambridge University Press.
Stocking, George. 1996. Franz Boas and the culture concept in historical per-
spective. American Anthropologist 68(4):867–882.
Sua´ rez, Jorge A. 1983. The Mesoamerican Indian languages. Cambridge: Cam-
bridge University Press.
Tomasello, Michael. 2008. Origins of human communication. Cambridge, MA:
Vilkuna, Maria. 1989. Free word order in Finnish: its syntax anddiscourse functions.
Helsinki: Finnish Literature Society.
Whorf, Benjamin L. 1956 (1945). Grammatical categories. In Language, thought,
and reality: selected writings of Benjamin Lee Whorf. John Carroll, ed. Pp. 87–
101. Cambridge, MA: MIT Press.
———. 1956 (1939). The relation of habitual thought and behaviorto language.
In Language, thought, and reality: selected writings of Benjamin Lee Whorf.John
Carroll, ed. Pp. 134–159. Cambridge, MA: MIT Press.
———. 1956 (1940). Science and linguistics. In Language, thought, and reality:
selected writings of Benjamin Lee Whorf. John Carroll, ed. Pp. 207–219. Cam-
bridge, MA: MIT Press.
Winawer, Jonathan, Nathan Witthoft, Michael C. Frank, Lisa Wu, Alex R. Wade,
and Lera Boroditsky. 2007. Russian blues reveal effects of language on color
discrimination. Proceedings of the National Academy of Sciences of the USA
Winford, Donald. 1993. Predication in Caribbean English creoles. Amsterdam:
Wittgenstein, Ludwig. 1953. Philosophical investigations. G. E. M. Anscombe,
trans. New York: Macmillan.
Wu, Ruey-Jiuan. 2004. Stance in talk: a conversation analysis of Mandarin ﬁnal
particles. Amsterdam: Benjamins. [JH]
Zinken, Jo¨ rg, and Eva Ogiermann. 2011. How to propose an action as anobjective
necessity: the case of Polish trzeba x (“one needs to x”). Research on Language
and Social Interaction 44(3):263–287.