Language Learning 44:3, September 1994,pp.

493-527

Review Article
Research on Negotiation: What Does It
Reveal About Second-Language Learning
Conditions, Processes, and Outcomes?
Teresa Pica
University of Pennsylvania

This article reviews insights into second-language (L2)

learning that have been revealed through over a decade of
research on the social interaction and negotiation of L2
learners and their interlocutors, begining with the seminal
work of Hatch (1978a, 197810) and Long (1980 et passim),
and withereferencet o a corpus of informal, experimental,
and classroom data from published studies. This research
illustrates ways in which negotiation contributes t o condi-
tions, processes, and outcomes ofL2 learningby facilitating
learners' comprehension and structural segmentation of
L2 input, access t o lexical form and'meaning, and produc-
tion of modified output. The research points out areas in
which negotiation does not appear to assist L2 learning,
especially with respect t o the learner's need to access L2

This article was written while the author was Ethel G. Carruth Associate
Professor of Education. A version was presented at the First Annual Pacific
Second Language Research Forum (PacSLRF), 14 July 1992, University of
Sydney, Australia. I acknowledge the many contributionsof my colleagues at
the University of Pennsyjvania, some of whom are now a t other universities,
who have helped me with collection and analysis of the data cited in this
paper.
Requests for reprints may be sent to the author a t Graduate School of
Education, University of Pennsylvania, 3700 Walnut Street, Philadelphia,
Pennsylvania 19104-6216. E-mail: billb8nwfs.gse.ubnn.edu

493
494 Language Learning Vol. 44, No. 3

grammatical morphology and t o strive to*ard accurate

productionaf L2 morphosyntax. Directions are suggested
for future research on negotiation and L2 learning through
longitudinal study, experimental’ design, and task-based
collection of data.

In 1978, Evelyn Hdtch made a pivotal and indelible mark on

the field of second langyage acquisition (SLA)through her publi-
cation oftwo seminal papers on language learning and interaction
(Hatch, 1978a, 1978b). With these papers, still widely cited, the
importance of interaction in language learning, once overlooked o r
at best taken for granted in SLA theory, became a major focus of
debate and discussion therein. By inviting and indeed Challenging
researchers t o look toward interaction for insight into second
lpguage (L2) development, Hatch called for a brand-new ap-
proach t o research. She encouraged a reversal of assumptions on
the nature ofthe learning process, as she urged researchers to turn
their attention away from questions about how L2 structure
learning led to the learner’s communicative use of L2, and instead
to examine how the learning of L2 structure euohed out of commu-
nicative use.
Fortunately, many researchers have continued t o take on
Hatch’s charge. Over the years, a fruitful, and often controversial,
line of research has evolved, much of it focused on a specific type
of interaction, which has come t o be known as negotiation. This
term has been used t o characterize the modification and restruc-
turing of interaction t h a t occurs when learners and their
interlocutors anticipate, perceive, or experience difficulties in
message comprehensibility. As they negotiate, they work linguis-
tically t o achieve the needed comprehensibility, whether repeating
a message verbatim, adjusting its syntax, chan’gingits words, b r
.
modifying its fov’m and meaning in a hast of other ways. .
Negotiation is not the only way’in which the interaction
between learners and their interlocutors can be modified or re-
structured. For example, the flow of their interaction can be
interrupted by a correction or rerouted t o a new topic. These two
 Pica 495

ways of modifying interaction have also been studied with respect

t o their possible contributions t o the learning process. But nego-
tiation, with its emphasis on achievingcomprehensibilityofmessage
meaning-both that provided to learners and that provided by
them-has sparked and sustained considerably more interest in
the field of SLA.
Of course, interest in language-learner interaction in general
and negotiation in particular did not begin with Hatch, nor with
SLA research for that matter. In a number ofthe incarnations that
the communicative approach to teaching has taken, methodolo-
gists have made the assumption that language learning takes
place through interaction, and have designed curricula accord-
ingly. (See reviews in Howatt, 1984; Kelly, 1969; Richards &
Rodgers, 1986;Yalden, 1983; and chapters throughout the recent
collection on task-based learning, Crookes & Gass, 1993a, 199313.)
In addition, sociolinguists have often looked at interaction
and negotiation among nonnative speakers (NNSs)and interlocu-
tors for what these processes revealed about the social roles and
responsibilities of this relationship (see Gumperz, 1964, 1970).
Indeed the term negotiation was used extensively in other fields
before its adoption by SLAresearchers,most ofwhom cite Garfinkle
(1967) as their primary source in this area (see, e.g., Scarcella &
Higa, 1981).
What made Hatch’s view so special, however, was her notion
that ‘interaction could also be used as a basis for examining the
linguistic and cognitive features of the L2 learning process, not
.just the social ones. She also made it clear.fiom the outset that
what was found could not necessarily be applied t o decisions about
classroom methods. We needed ti, “apply with caution,”as the title
of one of her early papers reminded us (Hatch, 1979). As Hatch was
careful t o point out, moreover, finding connections between inter-
action and SLA was easier’said than done. For one thing, a more
obvious role of the exchanges between learners and interlocutors
was for learners t o “check out the vocabulary” (Hatch, 1978b, p.
431) and to foreground background information; their role in
syntax building was not immediately apparent.’
496 Language Learn i.ng Vol. 44, No. 3

Further, Hatch (197813, pp. 432-433) noted that t h e work

needed t o draw o u t connections between interaction and SLA
coulci be messy, keeping researchers a t the level of observation,
impression, and anecdote. As she suggested, researchers might be
able to provide interesting samples of data, but such data could
lack the scope and consistency needed t o shed light on whether,
and if so how, language-learner discourse played a fundamental
role in language learning conhtions, processes, and outcomes.
Much observation, anecdote, and interpretation have indeed
emerged from the study of language-learner interaction. Fortu-
nately, a good deal of the anecdotal material has been compelling
and has contributed a wealth af descriptive data. These data,
together with those from some more focused studies, have contrib-
uted a good deal of information on how L2 learners2 interact
socially, linguistically, and ever so variably with native-speakers
(NSs)as well as with other learners. (See books by Allwright, 1988;
Chaudron, 1988; Ellis, 1985b; Larsen-Freeman & Long, 1991; van
Lier, 1988 for review chapters; the edited volumes of Day, 1986;
Fine, 1988; Gass & Madden, 1985; and recent state-of-the-art
papers by Long, 1992, and Wesche, in press.)
What might have been a messy task has actually been
handled with a great deal of consistency, responsiveness, and
responsibility by SLA researchers. They have developed theoreti-
cal models m d frameworks t o examine connections between the
properties of social discourse and the prQcesses of L2 learning, oRen
to test current SLA theories as well as &I propose new ones. These
studies have kept alive not only questions about the role of interac-
tion in the learning process, but also more h d a m e n t a l questions
about the learning process itself. (For illustrative, but by no means
exhaustive work, see, e.g., Chaudron, 1983a, 1983b, 1986; Crookes &
Rulon, 1985; Duff, 1986; Early, 1985; Ellis, 1985a, 1987; Gass &
Varonis, 1984, 1985a, 1985b, 1986, 1989; Lightbown, 1983; Long,
l980,1983,1985a, 1985b;Pica, 1991,1992;Pica, Doughty, &Young,
1986;Pica,Holliday,Lewis,Berducci, & Newman 1991;Pica, Holliday,
Lewis, & Morgenthaler 1989;Pica,Young, &Doughty, 1987a;Porter,
1986; Sato, 1986; Varonis & Gass, 1982,1985a, 1985b.l
 Pica 497

The seminal work of Michael Long has enriched and guided

this research all along. As early as 1980, Long investigated the
social dxcourse of NNSs and their NS interlocutors, and identified
the work that the NS and NNS do t o avoid and repair impasses in
their conversational discourse. At the time, he called this work
interactional modification (see Long, 1980, 1981, 1983). Later
Long and others also referred t o it as negotiation (see, e.g., Gass &
Varonis, 1985, 1986; Larsen-Freeman & Long, 1991, Scarcella &
Higa, 1981; Varonis & Gass, 1985a, 1985b).
Various labels have been applied t o its component features.
In his earlier writing, Long (1980)referred t o them as clarification
requests, confirmation checks, and comprehension checks, which
he then sorted into the broader categories of strategies, which.had
t o do with speakers’ utterance planning, and tactics, which related
t o utterance repair (Long, 1981). Yet another set of labels was
applied by Varonis and Gass (1985a, 1985b). They referred to
clarification requests and confirmation checks as indicators, pro-
duced because a preceding ‘utterarice had served to trigger them.
Later, Pica et al. (1989) and Pica et al. (1991) argued that labels
such a s clarification requests, confirmation checks, and indicators
implied that the researcher could identify a speaker’s intentional-
ity even though this was seldom the case. They therefore substituted
the term signals for those listener utterances in reaction to a
trigger utterance from the speaker. Varonis and Gass (1985a,
198513) also referred t o pushdowns and popups in the course of
learners’ negotiations, these having more t o do with turning points
in the overall structure of learners’ interactions than with utter-
ances directed toward message comprehensibility.
Whatever labels are used, these features of negotiation por-
tray a process in which a listenkr requests message clarification
and confirmation and a speaker follows up these requests, often
through repeating, elaborating, or simplifying the original mes-
sage. Such features, although prevalent throughout NS-NS social
discourse, are nevertheless significantly more abundant among
NS-NNS, even more so during NNS-NNS interaction (see Long,
1981; Varonis & Gass, 1985a, 1985b, respectively).
498 Language Learning Vol. 44, No. 3

Examples of NS-NNS negotiatiov from Long’s research are

provided below. Here, as well as in the excerpts to be discussed
later, the various checks, requests, and questions of negotiation
will be referred t o as signals. Shown in italics throughout the
excerpts, signals may be viewed a s utterances made by one
interlocutor t o the other that greater message comprehensibility
is needed. (Responses t o signals are shown in boldface). Such
needs may be apparent in the listener’s encoding of the signal,
through open-ended questions such as What did you say? o r X ?
What do you mean byX? or through requests for confirmation such
as Did you say X ? Such a need might also be inferred from the
speaker’s responses t o the listener’s signal, in which the speaker
repeats the original message for the listener, reduces o r increases
its length, substitutes new words, eliminates old ones, o r modifies
the message in a range of other ways.
In Example 1,for instance, all o r part ofthe NNS contribution
o f B u t uh but u h we take a break. . . break t i m e . . . thirty minutes
has served as a trigger for the NS t o signal at ten thirty you take a
break? Although this signal was not exactly what the NNS had
said initially, it did provide a context for the NNS t o respond, (as
shown in bold here and in the excerpts t o follow) by segmenting
and repeating part of the original message. This, in turn, led t o a
series of signal-response exchanges in which eventually, both the
time of the NNS B R E W and its duration were communicated
effectively.
1. NNS: But uh but uh . . we take we take a break. .
NS: oh
NNS: You know thirty minutes
NS: oh
NNS: Break time
NS: oh good
NNS: thirty minutes
NS: At ten thirty you take a break?
NNS: Thirty pinutes
NS: Right Wheq do you take the break? At ten
thirty?
NNS: Uhm . ten fifteen
 Pica 499

NS: Ten fifteen

NNS: ten fifteen From ten fifteen to ten fifty
five
NS: Ten forty-five
NNS: Ah ten f-forty-five
NS: Right right Have you seen Los Angeles?
Larsen-Freeman & Long, 1991, pp. 148)
In Example 2, the NNS signal, pardon me, brought forth an
NS response that elaborated and paraphrased much ofthe original
utterance. And in Example 3, the NNS signal what? what you say?
also brought forth a NS modlfied response, one characterized by
word and phrase substitution, in which by yourself was replaced
with alone, and the original reference t o here was made more
specific through to the states.
2. NNS: There has been a lot of talk lately about additives
and preservatives in food. In what ways has this
changed your eating habits?
NS: Uh, I don’t eat that many foods with preserva-
tives, anyway even before all the talk.
NNS: Pardon me?
NS: I don’t eat uh,canned foods or foods that
have preservatives.
(Gass & Varonis, 1985b, p. 50)
3. NS: so you came here by yourself or did you come with
friends?
NNS: no no I-what? what you say?
‘NS: did you come to the states with friends or
did you come alone?
NNS: no, alone-from Toronto (Pica, 1987, p. 5)
As these excerpts reveal, negotiation can serve a s a means of
working through perceived o r actual gaps in communication. Less
obvious are the ways in which it cin assist language learning. This
is perhaps because negotiation research has focused primarily on
language learning conditions rather than outcomes. Thus,re-
searchers have given less attention t o identifying a direct impact
for learners’ negotiation on restructuring of their interlanguage
grammar than t o documenting the contributions of negotiation in
bringing about conditions claimed t o be helpful for SLA, namely
500 Language Learn'ing Vol. 44, No. 3

learners' comprehension of L2 input, their production of modified

output, and their attention t o L2 form. As will be illustrated below,
in additional selections of negotiation and through statistical data
from the research from which these selections were taken, there is
a growing body of empirical evidence for the contributions of
negotigtion t o these three conditions, and a good deal oftheoretical
argument as well. This theoretical background is addressed in the
following section as a preliminary t o a review of the empirical
support.

Theoretical Perspectives on Conditions for SLA

The six theoretical perspectives on conditions for SLA dis-

cussed in this section fall into two broad categories. Three relate
t o what learners need to do t o learn a language and three are more
focused on what needs to be done with the L2 for it t o be learned.
Each will be addressed somewhat separately in this section, but
what will become more evident in later sections ofthis paper is that
the various conditions are h g h l y interrelated, both within and
across their respective categories.

Learner-Oriented Conditions

Ofthe first group of learner-oriented conditions, the one most

widely espoused is that: ( 1)comprehension of message meaning is
necessary if learners are t o internalize L2 forms and structures
that encode the message. Thus, exposure t o L2 input is not
sufficient for learners to be able tb access and internalize the L2
rules, forms, and features. This input must be made comprehen-
sible if it is t o assist the acquisition process (see, e.g., Krashen,
1980,1985; Long, 1980, l983,1985a, 1985b).
Adding t o this argument, Swain (1985) has claimed that: (2)
learner production of modifled output is also necessary for L2
mastery. She noted that it is possible to understand the mea&ng
of an utterance without reliance on or recognition of its morphol-
ogy or syntax; however, t o produce an utterance that can be
 Pica 501

understood often requires specific morphology and syntax to con-

vey its meaning. This, she speculated, might be why the spoken
production of Canahan immersion students has been found to lag
behind their listening comprehension. Even though, presumably,
these learners have been,given large amounts of comprehensible
input, this is not sufficient for the learning. Swain proposed
therefore that they might need t o be given opportunities to produce
comprehensible output; that is, t o organize and restructure their
, output syntactically. Swain thus views learners’ modification of
their output as a vehicle for them t o attend t o their interlanguage
grammar and thereby manipulate it in creqtive, complex, and
ultimately more target-like ways. (See Swain, 1985, for a more
detailed exposition of this argument.)
Both comprehension and modified production are tied closely
t o learners’ attention t o L2 form, with comprehension seen (par-
ticularly by Krashen, 1983, 1985) as the “entrance requirement”
for access t o form, and mbdified production as a context for
learners t o draw on their current system of interlanguage forms.
Indeed, learners’ awareness of form in L2 input is currently held
t o play a much more crucial role in successful SLA than was
assumed. This view is reflected in yet another theoretical claim
that: (3) attention t o L2 form is needed as learners attempt to
process meaningful inpuLUlong, 1990; Rutherford & Sharwood
Smith, 1985; Schmidt, 1990; Sharwood Smith, 1991) and attempt
t o master structural features that are difficult t o learn inductively
because they are relatively imperceptible in L2 input or overlap
with structures in the learner’s L1 (see Lightbown & Spada, 1990;
White, Spada, Lightbown, & Ranta, 1992).
The claim that L2 form must come to learners’ attention as
they process input also harkens back to Swain’s (1985) argument
regarding the importance of production a s a more direct means
than compreherision for learners to focus on form. For Swain,
learners’production of output,especially their modification thereof,
draws their attention t o the L2 structures that make it possible for
them t o convey message meaning. A more likely theoretical
variation, however, is that during output production learners’
502 Language Learning Vol. 44, No. 3

attention is focused on the form of their own interlanguage system

rather than the form of the L2 they wish t o acquire.

Language - 0r iented Conditions

The next set of claims about L2 learning conditions, as noted

above, centers on language processes or, more specifically, input
processes. The arguments here overlap a good deal with the
theoretical claims on learner-oriented conditions, especially with
I respect to their underlying assumptions and research.based They
simply offer another perspective on the same set of conditions for
L2 learning.
First among this second set of claims is the argument that: (1
positive L2 input-that is, input that is grammatically system-
atic-must be available t o serve the learning process. .SLA
theories all assume that learners draw on L2 input as data for their
learning. Any debate seems t o be over the nature of such data and
over whether and how the data must be organized for language
learning t o proceed-for example, t o allow learners t o imitate and
respond to linguistic stimuli, t o adjust innate structures, t o recog-
nize restrictions on lexis, t o discover rules through hypothesis
testing, and so forth.
Another language-centered claim is that: (2) enhanced L2
input, which makes subtle L2 features more salient for learners,
can assist their learning processes. Of parti'cular benefit is input
that provides information t o help learners identify which forms
can occur in the L2 and which cannot (Shanvood Smith, 1991).
Along similar lines, it is held that: (3)feedback and negative input
are needed t o provide learners with metalinguistic information on
the clarity, accuracy,and/or comprehensibilityoftheir interlanguage
(Schachter, 1983,1984; 1986,1991)and with structural information
that may help them notice non-target-likeforms-in their interlanguage
that are difficult to detect from positive input alone (Lightbown &
Spada, 1990; White, 1991; White et al., 1992). Both (2) and (3)are
variations on their more learner-centered version, (3) above, which
argues for learners'need to focus on L2 form to master the L2 system.
 Pica 503

To illustrate the relationship of negotiation to learners’ com-

prehension of L2 input, their production of modified output, &d
their access t o L2 form, the follbwing overview will present results
of several studies that have documented negotiation’s positive
effect on these processes.

Negotiation as an Aid to L2 Comprehension

There is considerable evidence for the role that negotiation

k
plays in bringing about compreh nsion, an important factor in
language learning. Indirect evide ce has come from studies such
as those by Chaudron (198333)and Long (1985a). Here, English L2
learners could better comprehend and recall ledurette content
when the words and sentences in the original versions of the
lecturettes were modified in ways that negotiation has been shown
t o generate. These modifications included repetitions-at the
word, phrase, and sentence level, the use of paraphrase, and the
insertion of conjunctions and enumerators t o mark relationships
of time and space.
Evidence of a .more direct relationship between negotiation
and comprehension has come from Pica et al. (1987a), who inves-
tigated the listening comprehension of 12 low-intermediate,
preacademic learners as they followed 30 different directions to a
picture assembly task under two input conditions. In one condi-
i
tion, the researchers first modified the direction uttera ces in
ways similar t o those used in the lecturette research described
above. Three femde NSs, who,were graduate students and
speakers of standard American English, then presented them
orally t o individual learners. Among the modifications in these
directions were reductions in syntactic complexity (from 1.20
clauses per sentence in the original directions to 1.02 in the
premodified) and repetitions and rephrasings of direction content
(from means of 2 0 content word repetitions and 16.47 total words
per each original direction t o means of 7.20 content word repeti-
tions and 33.47 total words for each premodilied direction).
In the other condition,the direction input was not premodified,
504 Language Learning Vol. 44, No. 3

but instead left in its original form, then presented by the same,
respective NSs during interaction with a matched group of learn-
ers. A s often a s possible, the NSs encouraged the learners to ask
for clarification of the directions (but not physical assistance with
carrying them out). An illustrative excerpt of what took place in
this second, negotiated input condition is shown in Example 4, and
will be discussed later in this section. Here, as noted in previous
excerpts, signals of negotiation a r e italicized. Responses t o these
signals a r e provided in boldface.
4. NS: [Initial, unmodified N S directions] Moving to the
top right corner, place the two mushrooms with
the three yellow dots in that grass patch down
toward the road. [Directionsduring negotiation]
Should I repeat it?
NNS: mmm
NS: OK. Moving to the top right corner, place
the two mushrooms with the three yellow
dots in that grass patch down toward the
road.
NNS: U m urhat means grass patch ?
NS: A piece of grass
NNS: Piece of grass, piece of grass
NS: Should I repeat it again? OK. Moving to the
top right corner, place the two mushrooms
with the three yellow dots in that grass
patch down toward the road.
NNS: I can’t understand
NS: OK. The top right corner of the picture.
NNS: mmm
NS: Place the two mushrooms, the two mush-
rooms with the three yellow dots in the
grass patch down towards the road
NNS: mm
NS: What’s the problem? Do you have a ques-
tion? I can help.
NNS: urn, on t h e - o n the right? left?
NS: On the top riglit corner of the picture
NNS: Top . . . top right right top corner
NS: in the grass patch down towards the road,
place the two mushrooms
 Pica 505

NNS: t w o mushrooms
NS: O K ? . . . All right.
(Pica, Young, & Doughty, 1987b)
As reported and discussed by Pica et al. (1987a), both the
premodified and the negotiated input conditions enhanced the
learners' comprehension, which for this study was operationalized
as the number of directions carried out accurately. However, the
negotiated input was significantly better in aiding direction com-
prehension. Thus, the mean comprehension score ofthe 6 learners
in the premodified input condition was 69% and that for the 6
learners in the interactionally modified condition was 88%(t=3.78,
p c . 0 5 ) . Further, a related, classroom-based study involving the
same assembly task found that, among two comparable classrooms
of learners, participants who witnessed their classmates negotiate
over the initially unmodified directions were able to comprehend
them as well as did participants who engaged in the actual negotia-
tion (Pica, 1991). Thus, when it comes to comprehension,negotiation
appears to be a powerful commodity; even learners' being allowed
only to observe negotiation can improve their comprehension.
Analyses of the language used during negotiation in these
studies uncovered the same kinds of manipulations of L2 l e i s and
structure that were made t o the premodified'directions. However,
these arose spontaneously in negotiation and there were signifi-
cantly more ofthem. Thus, content word repetition increased from
its initial mean of 0.20 per direction to a mean of 13.17 as a r e s d t
of negotiation (compared t o the increase t o 7.20 per premodified
directions). Further, the mean number oftotal words per direction
increased from 16.47 t o 51.64 (compared to an increase to 33.47 in
the premodified).
These negotiated modifications in direction length and re-
dundancy are typfied in Example 4 above. Here, content words
such as mushrooms, dots, andgrass were frequently repeated and
rephrased into longer stretches of direction input. These modifi-
cations were also part of a general phenomenon whereby the
direction input was broken down or segmentedinto more processible
units, which was particularly apparent when the learners asked
506 Language Learning Vol. 44, No. 3

open-ended questions. Typically, the NS responded. by segment-

ing the direction utterances into individual words and short
phrases, and repeating them or providing greater detail.
Thus, as illustrated ih Example 4, the NS took a direction
such as moving to the top right corner, place the two mushrooms
with the three yellow dots i n that grass patch down toward the road,
and extracted content words such as two mushrooms, three yellow
dots, and grass patch, as she defined them or offered a fuller
description t o assist the learner. When the learner asked about the
meaning ofgrass patch, for example, the NS extracted grass and
put it into the mbre comprehensible piece of grass.
This suggested that the input modificationg of negotiation,
although directed toward successful communication of the direc-
tion meaning, were oRen accomplished through segmentation of
the direction structure, done in ways that helped draw learners’
attention t o L2 form. This possible link between negotiation and
the processing of L2 form serves t o extend the work of negotiation
beyond that of helping learners with comprehension of message
meaning, and thus t o make negotiation a potentially powerful
contributor t o other dimensions of the L2 learning process. Such
contributions will be hscussed in the following section.

Negotiation, Comprehension, and Access to Form

The research findings reported above corroborate what has

been found from Long’s (1980)early work t o more recent research
by Holliday (1993) on the nature of input during negotiation: that
input modifications are significantly more abundant during nego-
tiation than during the rest of learners’ interaction. The findings
also provide further evidence that negotiation modifies the L2 in
ways that help learners comprehend its meaning. This perspec-
tive on negotiation i s highly restricted, however, and places
negotiation in a secondary role in L2 learning, because it sees
comprehension of meaning as the principal way t o access and
internalize L2 form, and negotiation as simply a way, albeit it a
very good way, t o bring comprehension about.
 Pica 507

Loolung at negotiation in this light makes negotiation impor-

tant for SLA only insofar as comprehension is important to SLA
This is unfortunate, because the role of comprehension in SLA has
become increasingly controversial. As a number of researchers
have argued (e.g., Chaudron, 1985; Ferch Jz Kasper, 1987; Gass,
1988; Shanvood Smith, 1987; White, 19881, it is difficult to find a
direct relationship between comprehension of meaning in L2 input
and the internalization of L2 forms that the learner aims to
acquire.
The most popular view on such a relationship is typified by
much of the work of Krashen (e.g., 1980, 1985). According to
Krashen, comprehensiop of meaning suffices t o enable learners to
access the forms and structures that encode that meaning. It is not
clear, however, what process makes this happen, and, a s a number
of researchers have pointed out, Krashen himself has not been’
consistent about this. (See Chaudron, 1985; Long, 1990; Schmidt,
1990;Schmidt & Frota, 1986; among others.) A s Chaudron (1985)
has noted, for example, Krashen (1983) proposed that learners
might access L2 form by noticing differences between forms in
their current level of competence and those in whatever input they
comprehended. In most of his other writing, however, Krashen
(1980, 1985) has said that the process of accessing form through
comprehension does not take place on so conscious a level. From
either of Krashen’s perspectives, however, the dmect connections
between comprehension of meaning and acquisition of form place
negotiation in a learning sequence in which it simply initiates L2
comprehension, whereas comprehension then serves a s the cata-
lyst for more direct L2 learning experiences. In such a sequence,
negotiation can lead t o comprehension of meaning, but it is
comprehension of meaning that leads t o a focus on, and eventually
acquisition of, L2 form.
A different sequence is equally likely, howe.ver, given the
great number of opportunities that can arise during negotiation to
draw learners’ attention to both message meaning and L2 form. In
this alternative sequence, the opportunities to hear a message
repeated, segmented, and reworded during negotiation are what
 Language Learning Vol. 44, No. 3

make it possible for the learner t o process the message and t o

comprehend its meaning. Negotiation data seem t o suggest,
therefore, that learners’ comprehension of meaning can be the
result of their access t o L2 form rather than its precursor.
A perspective on negotiation as directly influencing the learn-
ing of L2 structures and forms fits well with negotiation data. Too
much manipulation of form has been observed during the negoti-
ated process of making L2 input comprehensible t o learners (and,
as will be shown below, during learners’ attempts t o make their
output comprehensible t o NSs) t o afford negotiation only a second-
ary role in.the learning process. Further, as suggested in portions
of the excerpts thus far, a s weli as in some of those that follow,
learners’ and interlocutors’ negotiating does not always lead t o
their immediate comprehension of meaning but does get them to
manipulate form. This twofold potential of negotiation-to assist
L2 comprehension and draw attention t o L2 form-affords it a
more powerful role in L2 leaming than has been claimed s o far.
Other examp1,es of how negotiation can draw attention t o
form have been revealed in studies of learner-NS interaction in
which the focus is on both learners’ comprehension and their
production ofmessage meaning. Two areas of interest here are the
feedback on the learners’ message form that negotiation can
trigger and the modification in message form with which learners
and NSs can respond. Negotiation’s role in assisting learners’
access to form will be discussed in the following sections, as
excerpts from two additional studies on learner-NS interaction are
presented and analyzed.

Beyond Comprehension

That negotiation can help learners with assistance=morethan

comprehension of L2 input was shown in two ’studies on NS-
learner interaction, which have been described in detail in previous
papers (Pica et al., 1989, Pica et al., 19911. These studies focused
not only on the linguistic modifications in the input that the NSs
provided as they and learners engaged in information exchange
 Pica 509

tasks, but also on the feedback that the learners weregivenand the
modified output they produced on these tasdks. Illustrations ofthe
role played by negotiation with respect t o these three areas can be
found throughout the excerpts of negotiation below. These have
been taken from the 2,361negotiation utterances generated by the
learners and NSs who participated in the studies. Of these
negotiation utterances, 558 and 675 were NS signal and response
utterances, respectively, and 578 and 550 were respectively learner
signal and response utterances. The excerpts have appeared
elsewhere; as noted following each.
The participants were 32 low-intermediate, preacademic L2
English learners, all speakers of Japanese L1, and 32 NSs of
standard American English, including students, skilled workers,
and professionals. They worked as same-sex and cross-sex NS-
learner dyads on four oral communication tasks. One of the tasks
required one member of dyad t o draw, and then describe, a picture
for the other t o replicate. A'similar task required them to reverse
these roles for the same assignment. Another task engaged dyad?
of learners and NSs in replicating unseen, researcher-supplied
picture sequences, for which each held equally pertinent informa-
tion. A final task involved the dyads in exchanging opinions about
the possible contributions of these tasks t o their language learn-
ing.
For each task, the data reveal numerous opportunities for
leaners t o attend t o L2 form in their negotiations with the NSs.
For example, when the learners signaled difficulty in understand-
ing the NSs, the NSs often repeated and reformulated their
original utterance for the leahers. Similarly, when the NSs
signaled that they could not unaerstand the learners, the latter
oRen gave these signals back as L2 reformulations of their own
interlanguage utterances. Opportunities for the learners to at-
tend t o their own interlanguage form were also abundant: for
example, when portions of their interlanguage utterances were
repeated back t o them in the NSs' signals. Such opportunities also
arose when the learners signaled the NSs and responded to them,
particularly with signals that provided interlanguage versions of
510 Language Learning VOl. 44, No. 3

the NS utterances and with responses that modified their own

original productions. How this attention t o form was accom-
plished and the implications it had for L2 learning are discussed
in the following sections, as excerpts of the negotiations and
supportive statistical data are reviewed.

Negotiation and Input for Learning

Learners and interlocutors negotiate to repair breakdowns in

commbnication o r , ensure mutual comprehension of meaning.
Therefore, many of the speech modifications used t o make input
comprehensible can be considered lexical ones, in that words
unfamiliar t o the learner are repeated, replaced, or defined in
isolation from the longer segments in which they were initially
uttered.
As noted above, other modifications, however, are structural,
in the sense that portions of utterances are segmented and often
relocated t o another constituent position-for example, from ob-
ject in an initial utterance t o subject in a follow-up utterance.
These segments can be uttered either as new phrases that can
function on their own o r in conjunction with a paraphrase, sy-n-
onym, or other types of lexical modification. Thus, of the 675 NS
utterances of response t o learner signals in o u r data, 226 (33%)
were purely lexical, but 241 (36%)involved a modification t o the
initial structure.
Much of the time, the structural modification consisted of
simple segmentation of an initial utterance, so that a portion of an
utterance was extracted and repeated on its own. This pattern can
be seen in the NS response in Example 5, in which the NS extracted
the phrase a chimney on the left from his initial sentence. In
Example 6, there is both segmentation and relocation of a portion
of the NS initial utterance, as carport was moved from object of
have t o subject of is. The NS response utterance in Example 7
displays segmentation with additional structural change, here in
grammatical morphology for -s plural, as well as lexical modifica-
tion through description.
 Pica 511

5. NS: therF’s a chimney on the left

NNS: What?
NS: a chimney on the left (Pica, 1992a)
6. NS: you have a carport on the leR side?
.NNS: left side?
NS: the carport is on the left side.
(Pica, 1992a)
7. NS: and then another window over bushes I think
NNS: bush?
NS: sort of bush, I think it’s supposed to be a
bush (Pica, 1992a)
And in Example 8, the learner signal regarding three black
windows was interpreted (correctly) by the NS as the apparent
need to have garage segmented, moved from object of have t o
subject of is, and then defined and described.
8. NS: and I have a garage on the side with three little
black windows
NNS: three black windows?
NS: you know what a garage is?
NNS: no.
NS: um, it’s attached to the house. It’s a build-
ingattached to the house in whicbgou keep
your cars and called a garage, OK, so it
looks like a big house and a little house, but
they’re attached
NNS: Oh, it’s a small house.
NS: Uhuh
NNS: Uhuh, and black roof?
NS: Uhuh
NNS: Yeah, oh, maybe, let’s see, yeah, I understand.
(Pica, 1992a)
Again, in Example 9, the NS’s structural modification seg-
mented a portion of his original utterance and repositioned it idhis
response. As shown, the NS segmented hinges and moved. it from
its original position as object ofhast o subject position in the phrase
hinges hold it together. The meaning o f hinges was also supplied
&s the NS added hold it together. Here again, negotiation gave the
learner information about the meaning of hinges as well as its
structural possibilities as verb object and subject.
512 Language Learning Vol. 44, No. 3

9. NS: The door has hinges.

NNS: Hinges? I don't know what that means.
NS: Like hinges hold it together
NNS: Uhuh (Pica, 1993, p. 440)
And in Example 10, when the learner asked about tree, tree
was then segmented from the NS original message and reposi-
tioned in right dislocation in the N S response. The NS also gave
information about the meaning of pointed-it had to do with the
top of the tree.
10. NS: Is the rest of the tree pointed?
NNS: Tree?
NS: Is it pointed on top? the tree?
(Pica, 1992a)
Even when learners misanalyze NS input in'their signals, the
NS response can offer helpful segmentation. This can be seen in
Examples 11 and 12. In Example 11, for instance, the learner
asked about buvdaplate a s though it were a specific lexical item,
unknown to her. Through negotiation, however, she discovered
that buudaplate was actually a prepositional phrase whose mean-
ing she already knew. This was accomplished a$ the NS extracted
above from the original prepositional phrase above the plate. The
learner was then able t o echo back this origiial phrase and the NS
continued with a further description of the picture t o be drawn.
11. NS: I have a piece of toast with a small pat of butter
on it
NNS: hm hmm .
NS: and above the plate
NNS: what is buvdaplate?
NS: above
NNS: above the plate
NS: yeah not up as ifyou are sitting at the table
it would be farther away from you than the
plate
NNS: h m h m m
(Pica, 1992b, p. 225; Pica, 1993, p. 440)
And in Example 12, the learner asked about rectangular and
rectangle as well as square except. In all three signals, the NS
 Pica 513

described the features of a rectangle, attending to its meaning. In

response t o the signal square except, an incorrect reanalysis by the
learner, the NS also drew the learner’sattention to its form. He did
this by segmentingsquare except from his prior utterance. Hithen
repositioned except, using it to connect a square, as object of is in
the first sentence of his response, t o a square, as subject of has in
his following sentence.
12.’ NS it’s a rectangular bench
NNS: rectangular?
NS: yeah it’s in the shape of a rectangle with urn you
know a rectangle has two long sides and two
short sides
NNS: rectangle?
NS: re-rectangle it’s it’s like a square except you
you flatten it out
NNS: square except
NS: uh a rectangle is a square
NNS: uhuh
NS: except a square has four equal sides
NNS: yes
NS: a rectangle has two sides that are much longer
and two sides that are much shorter
NNS: OK (Pica, 1993,p. 437)
Whether or not the learner noticed these structural changes
cannot be determined from the data, but their frequent occurrence
suggests once again that negotiation as a process not only
emphasises communication of message meaning, it also provides
an opportunity t o focus on messqge form. Research on the role of
negotiation in supplying feedback,and in providing learners with
opportunities t o modify their output further substantiate this.

Negotiation as a Source of Feedback to Learners and a

Context for Their Modification of Output

That negotiation offers learners /more than assistance with

comprehension of L2 input is made quite clear when analysis
extends beyond the input learners are given to the output they
514 Language Learning Vol. 44, No. 3

produce. As noted above, when learners use their interlanguage

resources t o communicate with NSs, they set up a basis from which
they can be giKen feedback on their production-the meaningful-
ness of its content, the processibility of its form, or both. As a
result, the learners modify an original message toward greater
comprehensibility, often adjusting its form. Thus, through feed-
back, negotiation brings learners’ attention t o L2 versions of their
interlanguage utterances and heightens their awareness of their
own interlanguage system.

N S Signals as Feedback to Learners

In the current database, modification t o the sounds and
structures of learners’ interlanguage was found in 263 of the NS
signal utterances-47% of the total 558. Such feedback can be
found in Example 13, for instance, in which the NS asked for
confirmation of the learner’s production of big while segmenting it
from the learner’s original utterance and repronouncing it in a
more target-like way. Similarly, in Example 14, the NS gave the
learner crossed t o compare with closed, and in Example 15,the NS
asked about the learner’s verb drew, while using its more contex-
tually appropriate infinitive form.
13. NNS: this country like bik
NSL: big?
NNS: yeah (Pica, 1993, p. 440)
14. NNS: the windows are crozed
NS: ‘the windows have what?
NNS: closed
NS: crossed? I’mnot sure whsrt y.ou’re saying there-
NNS: windows are closed
NS: oh the windows are closed oh OK s o n y
(Pica, 1992a)
15. NNS: but I didn’t know how drew so we are very
confused
NS: to draw?
NNS: yeah (Pica et al., 1989, p. 89)

Other signal utterances contained more elaborate modifica-

 Pica 515

tions of form. In Example 16, for instance,bthe NS signal offered

the learner an L2 version that focused on differences in both form
and meaning by: segmenting tree from the learner’s utterance,
making it the subject of the NS’s sentence, modifying it with the
plural -s morpheme, and substituting branches for stick.
16. NNS: and tree with stick
NS: you mean the trees have branches?
NNS: yes (Pica, 1992a)

Learner Signals to NSs and

Segmentation o f L 2 Message Structure

The modifications that learners made in their signals and

responses to signals were similar to those in the signals and
responses of the NSs. As were the NSs’ output modifications, the
learners’ responses were distributed ’across lexis and structure,
with simple structural segmentation predominating over lexical
modification. This probablyreflects the learners’limited linguistic
resources for performing much modification beyond extracting
recognizable forms from the NS input.
Thus, of the 578 learner signals to NSs, 352 (61%)involved
simple structural segmentation as shown in Examples 17,18,and
19 below. Most of the other signals (13%)were either open
questions, such as What?,which did not involve modification, or
lexically modified versions of what the NS had said (9%). Taken
together, these data indicated that learner signals focused prima-
rily on the structure of NS utterances, and that in signalling about
NS mesage meaning, the learners were’particularly sensitive t o
smaller segmefits of NS input.
17. NS: are they facing one,another?
NNS: Facing?
NS: urn are the chairs at opposite ‘endsof the
table or-
NNS: yeah (Pica, 1993, p. 440)
18. NS: you have a carport on the left side?
NNS: left side?
516 Language Learning Vol. 44, No. 3

NS: the carport is on t h e left side.

(Pica, 1992a)
19. NS: we got a plant.
NNS: plant?
NS: yeah, urn it’s kind of like a fern, h a s a lot of big
leaves; it’s in a pot. (Pica, 1992a)

Learner Responses to N S s as Production of Modified Output

Analysis of the data revealed that many learner responsesto

N S signals also contained some kind of structbral modification.
Typically, what learners did was t o segment portions of their
initial utterances and use them to form their responses. Thus, of
the 550 learner responses, 148 (27%) involved structural modifica-
tion, and of these structurally modified utterances 83 (56%) were
characterized by simple segmentation o r by segmentation accom-
panied by anothel: type of linguistic modification. This was shown
in Example 14 above, in which the N S requested clarification ofthe
learner crozed by repeating most ofthe learner initial message, but
inserting what? for the part of the message he could not under-
stand. The learner was also given crossed t o compare with closed.
That the learner’s attention was focused on this form was shown
as he segmented windows from the initial utterance, but then
incorporated the repronounced version of closed into his response.
In distinction t o the frequency and consistency of modifica-
tion in NS responses t o learner signal$, the m’odification in learner
responses was less frequent and was influenced considerably by
the type of signal used by the NS. Thus, NS signals drew responses
from learners that also m o a e d message meaning and form, but
this was twice as likely t o occur if the NS signal was an’open
question such a s What? or huh? or Can you tell me what that
means? than if it was a repetition o r modification of the learner’s
message. This can be seen, for instance, in Example 20, in which
the learner elaborated about patton (albeit unsuccessfully) in
response t o the NS open signal.
20. NNS: we have common patton in this case
 Pica 517

NS: I don’t know that word.. . can you describe what

it means
NNS: yes uh uh if I can explain the car‘s nature,
we understand easy because car has a few-
a lot of nature.. . (Pica, 1989, p. 88)
A very different, but more typical type of response was shown
earlier in Examples 13, 15, and 16, in which ihe learners simply
said yeah or yes t o the N S interpretation of their message, rather
than attempt their own. Thus, modification in learners’ responses
was not inevitable, but instead dependent on the type of NS signal
used. This differed from the somewhat unrestrained use of
modification by NSs responding t o learner signals, such as in
Examples 6,7,10,17,18, and 19. (See Pica, 1992; Pica, Lewis, &
Holliday, 1990 for a fuller discussion; and Pica, Lincoln-Porter,
Paninos, & Linnell, in pyeparation, for supportive, more focused
research. 1

Negotiation and SLA: Research Issues and Directions

Given negotiation’s many apparent contributions to L2 learn-

ing, i t may seem difficult t o understand why we do not take greater
advantage of negotiation and really push for it-as teachers in o u r
classrooms, as SLA researchers, and in o u r own experiences as
language learners. Yet there are many reasons, both methodologi-
cal and social, not to do so.
First, negotiation must be looked a t candidly. The emphasis
in this paper so far has been on what negotiation can do for L2
learning; it is important t o look as well at what negotiation has not
been able t o do. The recent proliferation of writing on SLA theory
helps researchers begin t o understand why no one experience,
activity, or endowment can account for all of L2 learning. On this
basis alone, negotiation cannot be really counted on, any more
than anything else can be counted on, as the be-all and end-all of
L2 learning.
This aside, a more specific explanation lies in negotiation’s
focus. Negotiation, by definition, focuses on the comprehensibility
518 Language Learning Vol. 44, No. 3

of message meaning, and on the message's form only insofar a s

that can contribute to its comprehensibility. - Learners and their
interlocutors find ways t o communicate messages through nego-
tiation, but not necessarily with target-like forms.
Further, negotiation seems t o work most readily on lexical
items and larger syntactic units. Negotiation over grammatical
morphology is rare, a t least on the various communication tasks
that have been studied t o date (Pica, Kanagy, & Falodun, 1993).
Even asking learners and their interlocutors t o tell stories, se-
quence events, or explain procedures in their tasks does not get
them to negotiate much over time and aspect marking. Instead,
they give greater attention t o the people in their pictures and
stories-what they look like, their shapes, sizes, and s o forth-
than t o what the people are doing. So the learners'segment and
move larger units of syntax such as sentence elements, for ex-
ample, but do little else. These findings do not mean that learners
and interlocutors cannot negotiate over verb tense and aspect, but
that many of the communication activities in which they partici-
pate-both in research and in everyday life-do not demand their
attention t o these areas of grammar.
Certainly learner variables come into play. As Pienemann
(1989) has shown, if learners are not ready for a new word, form,
or rule, they cannot acquire it, and thus negotiation will not help
toward its internalization. The interplay of L2 code and discourse
also plays a role. As Sat0 (1986) has argued and shown, discourse
constraints are often such that learners do not need much
morphosyntax t o communicate with their interlocutors. Further,
many features of language used in communication are barely
detectable, impossible for learners to attend to. This may be why
researchers such as White et al. (1992) have been successful in
assisting SLA through enhanced L2 input, highlighting relatively
imperceptible linguistic units, complex rules, or features that are
difficult for learners t o differentiate from those in their L1.
Negotiation must also be put into critical perspective from a
social point of view. Although common to many aspects of social
discourse, negotiation needs t o be balanced by the smooth flow of
 Pica 519

even exchange. As Aston (1986) has argued, too many impasses

and repairs can make for uneasy social relationships. Too many
clarification questions can be downright annoying.
Besides, no one yet agrees on how t o measure the impact of
negotiation on the internalization of L2 knowledge: Should re-
searchers use grammaticality judgments? L2 use in context?
spontaneous use in communication? change over time? Before
casting aside negotiation as a key factor in the learning process,
researchers need t o agree on what constitutes L2 learning. Most
research has taken a process approach toward characterizing L2
learning through negotiation, but if negotiation’s role in learning
is t o be tested more fully, an outcome approach will be necessary
as well.
Related t o the matter of specifying learning outcomes are
matters of research design. SLA researchers must ask whether
they have been investigating negotiation in ways that tap into its
total potential for assisting L2 learning. We cannot expect defini-
tive answers from naturalistic studies that focus on interaction in
a variety of learning contexts but not negotiation per se. Yet there‘
is little else t o turn to. Longitudinal research by Sat0 (1986) and
Schmidt and Frota (1986),for example, has shed much light on L2
learning processes. However, because their students participated
in negotiation as one of many different types of interaction in their
daily lives, one cannot identify the i n f l u e n c w r lack thereof-
that negotiation might have had on their learning.
Especially important, two experimental studies, focused
largely on negotiation, have shown i t has positive, short-term
effects on L2 learning. Doughty’s (1988,1992)work on English L2
relative clause learning and Loschky’s (1989) on locatives in
Japanese have used negotiation t o provide an inundation of these
formal structures t o L2 learners. Their positive results for learn-
ing outcomes reveal a great deal about what negotiation can do
when put t o the test.
In Doughty’s (1988, 1992) research, deliberate and height-
ened contexts for negotiation enabled learners t o be given modified
input that was rich in relative clauses. This may be the very route
520 Language Learning Vol. 44, No. 3

that must be taken in studying negotiation’s effect on learning.

Such treatment may not seem “natural”or socially acceptable, but
we should not overlook how highly effective it was for Doughty’s
participants in the short term. We still have little idea of what
intensive negotiation can do for learning outcomes, especially if
the learner were ready t o learn a particular L2 feature. But this
certainly deserves more s t u h e s than Doughty’s promising contri-
bution.
In addition, i t increasingly appears that experimental stud-
ies may best lead t o important insights into what negotiation can
do for L2 learning. With respect t o analyses of negotiation data,
tremendous challenges lie ahead. A s noted earlier, these data
show how learners’ attention cari be brought t o differences be-
tween their own production and a target model. However, we must
emphasize can, because negotiation data seldom obviously show
whether o r not learners perceive these differences. Overall, we can
safely say that negotiation provides learriers with opportunities t o
atteqd t o L2 form and t o relationships of form and meaning.
Whether they indeed do cannot be observed, o r even inferred, most
of the time. This may not reflect a fault of negotiation, but rather
the current state of research instruments:
So where is research with respect t o language learning
through negotiation? Again, thanks to Hatch (1978a, 1978b, 1979)
andLong(1980,1981,1983,1985a,1985b,1990,1992)andtherich
line of research they inspired, we know that negotiation, as a
particular way of modifying interaction, can accomplish a great
deal for SLA. It can help make input comprehensible t o learners,
help them modify their own output, and provide opportunities for
them t o access L2 form and meaning. However, we need greater
insight into how i t is connected t o learning outcomes.
We also badly need more experimental studies comparing the
effects on L2 learning outcomes ofinteraction with negotiation and
that without negotiation. These studies should use tasks designed
t o tap into grammatical modifications-in itself, not so easy. They
then should be followed up with long-term, very focused experi-
mental studies: also not an easy chore, and one with its own
 Pica

logistical and financial problems. Perhaps such research can be

done in classrooms, using approaches similarto those oflightbown,
Spada, and White (see Lightbown & Spada, 1990; White, 1991;
White et al., 1992). This, however, will require a different sort of
classroom than researchers and research consumers have been
used to.
Res,earch on classroom interaction has already shown why
there is s o little negotiation in the language classroom context,
much of it related t o matters of teacher and student power, to
trgditions in language teaching, and t o expectations about the
language classroom (see discussion in Pica, 1987). There is also
the possibility that the drive t o promote in the classroom discourse
which resembles that found outside it has provided learners with
tasks that get them t o talk t o their teacher and t o each other but
not t o ask the kinds of questions about form and meaning that will
really deliver L2 data t o them. Clearly, newer tasks are needed for
both classroom learning and classroom research. (See Loschky &
Bley-Vroman, 1990, 1993; Crookes & Gass 1993a, 1993b, for
promising new directions in this area.) Any progress on these
fronts will accomplish a great deal not only for SLA research, but
also for classroom methods, and for the learners whom they are
intended to serve. Much work still needs to be done on the study
of language learning through negotiation, but it is work that will
sustain the liveliness and curiosity of those who participate in the
field of SLA as teachers, learners, and researchers.
Revised version accepted 23 November 1993

Notes
'I thank an anonymous reviewer for pointing this out to me.
21n this article, the terms learner and NNS reflect the ways in which the
researchers cited have characterized their participants. However, the use of
the term learner by the present author reflects her current perspective on the
participants in her studies, as L2 learners who are also NNSs of English L2.
522 Language Learning Vol. 44, No. 3

References

Allwright, R. (1988). Observation in thelanguageclassroom. London: Longman.

Aston, G. (1986). Trouble shootingin interaction with learners: The more the
merrier? Applied Linguistics, 7, 128-143.
Chaudron, C. (1983). Foreigner talk in the classroom-An aid tolearning? In
H. Seliger & M. Long (Eds.), Classroom-oriented research in second
language acquisition (pp. 127-143). Rowley, MA: Newbury House.
Chaudron, C. (1983). Simplification of input: Topic reinstatements and their
effect on L2 learners’recognition and recall. TESOL Quarterly, 17,437458.
Chaudron, C. 1985. Intake: On models and methods for discovering learners’
processing of input. Studies in Second Language Acquisition, 7, 1-14.
Chaudron, C. (1986). Teachers’ priorities in correcting learners’ errors in
French immersion classes. In R. R. Day (Ed.), Talking to learn: Conversa-
tion in second language acquisition (pp. 64-84). Rowley, k4: Newbury.
Chaudron, C. (1988). Second language classrooms: Research on teaching and
learning. New York: Cambridge University Press.
Crookes, G., & Gass, S. M. (Eds.). (1993). Tasks and language learning:
Zntegrating theory and practice. Clevedon, Avon: Multilingual Matters.
Crookes, G., & Gass, S. M. (Eds.). (1993). Tasks in a pedagogical context:
Integrating theory and practice. Clevedon, Avon: Multilingual Matters. ’
Crookes, G., & Rulon, K. (1985). Incorporation ofcorrective feedback in native
speakerlnon-native speaker conversation (Tech. Rep. No. 3). The Center
for Second Language Classroom Research/Social Science Research Insti-
tute, University of Hawaii at Manoa.
Day, R. R. (Ed.). (1986). Talking to learn: Conversation i n second language
acquisition. Rowley, MA: Newbury House.
Doughty, C. (1988).An empirical study ofrelatiuization in English as a second
language. Unpublished doctoral dissertation, University of Pennsylvania,
Philadelphia.
Doughty, C. (1992). Second language instruction does make a difference:
Evidence from a n empirical study of SL relativizntion. Studies in Second
Language Acquisition, 13, 431470.
Duff, P. (1986). Another look at interlanguage talk: Taking task to task. In R.
R, Day (Ed.). Talking to Learn: Conversation in second language acquisi-
tion (pp. 147-181). Rowley, MA: Newbury House.
Early, M. (1985). Input and interaction in content classrooms: Foreigner talk
and teacher talk in classroom discourse. Unpublished doctoral disserta-
tion, University of California at Los Angeles.
Ellis, R. (1985a). Teacher-pupil interaction in second language development.
In S. M. Gass & C. G. Madden (Eds.), Input in second language acquisition
(pp. 68-88). Rowley, MA: Newbury House.
 Pica 523

Ellis, R. (1985b). Understanding second fdnguage acquisition. Oxford: O W .

Ellis, R. (1987). Second languageacquisition in context. Englewood Cliffs, NJ:
Prentice-Hall.
Fzerch, C., & G Kasper. (1987). The role ofcomprehension in second language
acquisition. Applied Linguistics, 8, 256-274.
Fine, J. (Ed). (1988). Second language discourse: A textbook of current
research. Norwood, NJ: Ablex.
Garfinklei H. (1967). Studies in ethnornethodology. Englewood Cliffs, N J
Prentice-Hall.
Gass, S. M. (1988). Integrating research areas: A framework for second
language studies. Applied Linguistics, 198-217.
B:
Gass, S . M., & Madden, C. G. (Eds.). (1985). Input in second language
I
acquisition. Rowlky, M A Newbury House.
Gass, S. M., & Varonis, E. (1984). The effect of familiarity on the
comprehensiblity of non-native speech. Language Learning, 34,65-89.
Gass, S . M., & Varonis, E. (1985a). Task variation and NNS/"S negotiation
ofmeaning. In S. M. Gass & C. G. Madden (Eds.), Input in second language
acquisition (pp. 149-161). Rowley, MA: Newbury House.
Gass, S. M., & Varonis, E. M. (1985b). Variation in native speaker speech
modification to non-native speakers. Studies in Second Language Acqui-
sition, 7, 37-58.
Gass, S. M., & Varonis, E. (1986). Sex differences in NNS/"S interactions.
In R. R. Day (Ed.). Talking to learn: Conversation in second Language
acquisition (pp. 327-4511. Rowley, M A Newbury House.
Gass, S. M., & Varonis, E. (1989). Incorporated repairs in NNS discourse. In
M. Eisenstein (Ed.), The dynamic interlanguage (pp. 71-86). New York:
Plenum.
Gumperz ,J . (1964.) Linguistic and social interaction in two communities. In
J. Gumperz & D. Hymes (Eds.), The ethnography of communication.
American Anthropologist, 66,6, part 11,137-154.
Gumperz , J. (1970). Verbal strategies in multilingual communication. In J.
Alatis (Ed.), Report of the Twenty-first Annual Round Table Meeting on
Linguistics and Language Studies (Monograph Series on Languages and
Linguistics). Reprinted in R. Abrahams & R. Troike (Eds.), (1972). Lan-
guage, Culture, and Education (pp. 184-196). Englewood Cliffs, NJ:
Pren tice-Hall .
Hatch, E. (1978a). Acquisition of syntax in a second language. In J. C .
Richards (Ed.), Understanding second and foreign language learning:
Issues an d approaches (pp. 34-69). Rowley, MA: Newbury House.
Hatch, E. (1978b). Discourse analysis and second language acquisition. In E.
Hatch (Ed.), Second language acquisition: A book of readings (pp. 401-
435). Rowley, MA. Newbury House.
524 Language Learning VOL.44, No. 3

Hatch, E. (1979). Apply with caution. Studies in Second Language Acquisi-

tion, 2, 123-143.
Holliday, L. (1993). N S syntactic modifications in N S - N N S negotiation as
input data for second language acquisition ofsyntax. Unpublished doctoral
dissertation, University of Pennsylvania, Phiadelphia.
Howatt, A. (1984). A history of English language teaching. Oxford: Oxford
University Press.
Kelly, I. (1969). Twenty-five centuries of language teaching. Rowley, Mk.
Newbury House.
Krashen, S. D. (1980). Second language acquisition and second language
learning. Oxford: Pergamon Press.
Krashen, S. D:(1983). Newmark’s Ignorance Hypothesis and current second
language acquisition theory. In S. M. Gass & L. Selinker (Eds.),Language
transfer in language learninglpp. 135-156). Rowley, MA: Newbury House.
Krashen, S. D. (1985). The Input Hypothesis: Issues and implications. Lon-
don: Longman.
Larsen-Freeman, D. E., & Long, M. (1991). A n introduction to second lan-
guage acquisition research: New York: Longman.
Lightbown, P. (1983). Exploring relationships between developmental and
instructional sequences in L2 acquisition. In H. Seliger & M. Long (Eds.),
Classroom-orientedresearchin second languageacquisition (pp. 217-343).
Rowley, MA: Newbury House.
Lightbown, P., & Spada, N. (1990). Focus-on-form and corrective feedback in
communicative language teaching: Effects on second language learning.
Studies i n Second Language Acquisition, 12; 429448.
Long, M. H. (1980). Input, interaction, and second language acquisition.
Unpublished doctoral dissertation. University of California, Los Angeles.
Long, M. H. (1981). Input,interaction,and second-language acquisition. In H.
Winitz (Ed.), Annals of the New York Academy of Sciences: Vol. 379.Native
language and foreign language acquisiton (pp. 259-278). New York: New
York Academy of Sciences.
Long, M. H. (1983).Linguistic and conversational adjustments to nonnative
speakers. Studies in Second Language Acquisition, 5, 177-193.
Long, M. H. (1985a). Input and second language acquisition theory. In S. M.
Gass & C. G. Madden (Eds.), Znput in second language acquisition (pp.
377-393). Rowley, MA: Newbury House.
Long, M. H. (198513).A role for instruction in second language acquisition:
Task-based language training. In K. Hyltenstam & M. Pienemann (Eds.)
Modelling and assessing second language acquisition (pp. 77-100).
Clevedo;, Avon: Multilingual Matters.
Long, M. H. (1990). The least a second language acquisition theory needs to
explain. TESOL Quarterly, 24, 649-666.
 Pica 525

Long, M. H. (1992, March). Input and focus onform. Paper presentedatAAAL

Convention, Seattle, WA.
Loschky, L. (1989). Negotiated interaction: The roles of task and culture.
Unpublished master’s thesis, University of Hawai’i, Honolulu.
Loschky, L., & Bley-Vroman, R. (1990). Creating structure-based tasks for
second language development. University of Hawai‘i Working Papers in
ESL; 9 , 161-212.
Loschky, L., & Bley-Vroman, R.. (1993). Grammar and task based methodol-
ogy. In G. Crookes & S. M. Gass (Eds.), Tasks and language learning:
Integrating theory and practice (pp. 123-167). Clevedon, Avon: Multilin-
gual Matters.
Rca, T. (1987). Second language acquisition, social interaction, and the
classroom. Applied Linguistics, 8, 3-21.
Rca, T. (1991). Classroom interaction, participation, & comprehension:
Redefining relationships. System, 19,437452.
Pica, T. (1992% July). Second language learning through intemciton and the
negotiaiton ofconditions,processes, and outcomes. Paper presented at first
Pacific Second Language Research Forum, Sydney, NSW, Australia.
Pica, T. (1992b). The textual outcomes of native speaker-non-native speaker
negotiation. In C. Kramsch &: S. McConnell-Ginet (Eds.), Text andcontext:

I
Cross-discipli ary perspectives on language study (pp. 198-137). Lexing-
ton, MA:D.C. Heath & Co.
Pica, T. (1993).C mmunication with second language learners: What does i t
reveal about t e social and linguistic processes of second language acqui-
sition? In J. Alatis (Ed.), Language, communication, and social meaning
(pp. 434-464). Washington, DC: Georgetown University Press.
Plca, T., Doughty, C., & Young, R. (1986). Making input comprehensible: Do
interactional modifications help? I T L Review ofApplied Linguistics, 72,l-
25.
Pica, T.,Holliday, L., Lewis, N, Berducci, D., & Newman, J.(1991).Language
learning through interaction: What role does gender play? Studies in
Second Language Acquisition, 13, 343-376.
Pica, T., Holliday, L., Lewis, N., & Morgenthaler, L. (1989). Comprehensible
output as an outcome of linguistic demands on the learner. Studies in
Second Language Acquisition, 11, 63-90.
Pica, T.,Kanagy, R., & Falodun, J. (1993). Choosing and using communica-
tion tasks for second language research and instruction. In S. M. Gass &.
G. Crookes (Eds.), Task-based learning in a second language (pp. 9-34).
Clevedon, Avon: Multilingual Matters.
Pica, T., Lewis, N., & Holliday, L. (1990, March). NS-NNS negotiation: An
equal opportunity for speech modification? Paper presented to 24th annual
TESOL Convention, San F’rancisco, CA.
526 Language Learning Vol. 44, No.,3

Pica, T., Lincoln-Porter, F., Paninos, D., & Linnell, J. (in preparation). Can
second language learners be a resource for their L2 learning? Only their
researcher knows for sure.
Pica, T., Young, R., & Doughty, C. (1987a). The impact of interaction on
comprehensiori? T E m L Quarterly, 21, 737-758.
Pica, T., Young, R., & Doughty, C. (1987b). [Transcription]. Unpublished raw
data.
Pienemann, M. (1989). Is language teachable?: Psycholinguistic experiments
and hypotheses. Applied Linguistics, 10, 52-79.
Porter, P. (1986). How learners talk to each other: Input and interaction in
task-centered discussions. In R. R. Day (Ed.), Talking to learn :Conversa-
tion in second language acquisition (pp. 200-224). Rowley, W Newbury
House.
Richards, J., & Rodgers, T. (1986). Approaches and methods in language
teaching. Cambridge: Cambridge University Press.
Rutherford, W., & Sharwood Smith, M. (1985). Consciousness-raising and
universal grammar. In W. Rutherford & M. Sharwood Smith (Eds.),
Grammar and second language learning (pp. 107-116). Rowley, MA:
Newbury House.
Sato, C. (1986). Conversation and interlanguage development: Rethinking
the connection. In R. R. Day (Ed.), Talking to learn: Conversation in second
language acquisition (pp. 2 3 4 8 ) . Rowley, MA: Newbury House.
Scarcella, R. C,, & Higa, C. (1981).Input, negotiation, and age differences in
second language acquisition. Language Learning, 3 1 , 4 0 9 4 3 7 .
Schachter, J. (1983).Nutritional needs of language learners. In M. Clarke &
J. Handscombe (Eds.), On TESOL '82: Pacific perspectives on language
learning and teaching (pp. 175-89). Washington, DC: TESOL.
Schachter, J. (1984). A universal input condition. In W. Rutherford (Ed.),
Universals and second language acquisition (pp. 167-183). Amsterdam:
John Benjamins.
Schachter, J. (1986). Three approaches to the study of input. Language
Learning, 36,211-225.
Schachter, J . (1991). Corrective feedback in historical perspective. Second
Langucrge Research, 7,89-102.
Schmidt, R. (1990). 'The role of consciousness in second language acquisition.
Applied Linguistics, 1 I, 129-158.
Schmidt, R., & Frota, S. (1986). Developing basic conversational ability in a
second language: A case study of an adult learner of Portuguese. In R. R.
Day (Ed.), Talking to learn: Conversation in second language acquisition
(pp. 239-326). Rowley, M A Newbury House.
Sharwood Smith, M. (1987). Comprehension vs. acquisition: Two ways of
processing input. Applied Linguistics, 8, 237-255.
 Pica 527

Sharwood Smith, M. (1991). Speaking to many minds: On the relevance of

different types of language information for the L2 learner. Second Lun-
guage Research, 7, 118-132.
Swain, M. K. (1985). Communicative competence: Some roles of comprehen-
sible input and comprehensible output in its development. In S. M. Gass
& C. G.Madden (Eds.),Input in second languageacquisition(pp.236-244).
Rowley, MA: Newbury House.
van Lier, L. (1988). The classroom and the language learner. New York:
Longman.
Varonis, E., & G a s , S. M.(1982). Thecomprehensibilityofnon-nativespeech.
Studies in Second Language Acquisition, 4,41-52.
Varonis, E., & Gass, S. M. (1985a). Miscommunication in nativehon-native
conversation. Language in Society, 14, 327-343.
Varonis, E., & Gass,S. M. (1985b). Non-nativehon-native conversations: a
model for the negotiation of meaning. Applied Linguistics, 6, 71-90.
Wesche, M. (in press). Input, interaction, and acquisition: The linguistic
environment of the language learner. In R. Richards & C. Gallaway (Eds.)
Input and interaction in language acquisition. Cambridge: Cambridge
University Press.
White, L. (1988).Againstcomprehensibleinput:Theinputhypothesisand the
development of second language competence. Applied Linguistics, 9,89-
110.
White, L. (1991). Adverb placement in second language acquisition: Some
effects of positive and negative evidence in the classroom. Second Lun-
. guage Research, 7, 133-161.
White, L., Spada, N., Lightbown, P., & Ranta, L. (1992). Input enhancement
and L2 question formation. Applied Linguistics, 12,416-432.
Yalden, J. (1983).The communicative syllabus: Evolution, design, and imple-
mentation. Oxford: Pergamon;