• unlimited access with print and download
    $ 37 00
  • read full document, no print or download, expires after 72 hours
    $ 4 99
More info
Unlimited access including download and printing, plus availability for reading and annotating in your in your Udini library.
  • Access to this article in your Udini library for 72 hours from purchase.
  • The article will not be available for download or print.
  • Upgrade to the full version of this document at a reduced price.
  • Your trial access payment is credited when purchasing the full version.
Buy
Continue searching

Linguistic alignment: The role of social and cognitive variables

ProQuest Dissertations and Theses, 2011
Dissertation
Author: Jacqueline M Coyle
Abstract:
The current study uses two experiments to explore how a range of social and cognitive variables affect linguistic alignment. Experiment 1 examines the effects of a number of individual difference variables (inhibitory control, pattern learning, creativity, trait conformity, and the need to belong), while Experiment 2 focuses on one specific social variable--female fertility. In order to look at linguistic alignment (or a lack thereof), the study employs a picture description task in which dyads (each dyad is comprised of one participant and one confederate) take turns describing pictures to each other. Both experiments measured the degree to which participants aligned with confederates on sentence structure, and Experiment 2 also measured alignment of vocal pitch within the dyads. The results of Experiment 1 suggest that individual differences do not affect the amount of alignment of sentence structure. However, the results of Experiment 2 reveal that the fertility level of a female conversational partner affects the degree to which men align with women on sentence structure. Specifically, men were less likely to align with the sentence structure of a woman who was at high fertility than with the sentence structure of a woman who was at a lower level of fertility. Experiment 2 also demonstrates that men and women in conversation do align on vocal pitch. Possible explanations and implications are discussed.

TABLE OF CONTENTS

List of Tables ............................................................................................................................... vii List of Figures ............................................................................................................................. viii Abstract ......................................................................................................................................... ix

1. INTRODUCTION .....................................................................................................................1

2. EXPERIMENT 1 .......................................................................................................................5 2.1 Inhibitory Control ..............................................................................................................5 2.2 Pattern Learning .................................................................................................................6 2.3 Creativity ............................................................................................................................6 2.4 Conformity & the Need to Belong .....................................................................................6 2.5 Method ...............................................................................................................................8 2.5.1 Participants ..............................................................................................................8 2.5.2 Confederates & Experimenters ...............................................................................8 2.5.3 Materials ..................................................................................................................8 2.5.4 Procedure ...............................................................................................................12 2.5.5 Design & Analysis .................................................................................................13 2.6 Results & Discussion .......................................................................................................14

3. EXPERIMENT 2 ......................................................................................................................17 3.1 Method .............................................................................................................................20 3.1.1 Participants ............................................................................................................20 3.1.2 Confederates ..........................................................................................................20 3.1.3 Experimenters ........................................................................................................21 3.1.4 Materials ................................................................................................................21 3.1.5 Procedure ...............................................................................................................23 3.1.6 Design & Analysis .................................................................................................23 3.2 Results & Discussion .......................................................................................................24 3.2.1 Structural Priming .................................................................................................24 3.2.2 Vocal Pitch ............................................................................................................32

4. GENERAL DISCUSSION ......................................................................................................35 4.1 Conclusion .......................................................................................................................38

APPENDICES ..............................................................................................................................40 A. FOOTNOTE .............................................................................................................................40 B. IRB APPROVAL LETTER .....................................................................................................42

vi

C. INFORMED CONSENT .........................................................................................................44

REFERENCES .............................................................................................................................46

BIOGRAPHICAL SKETCH ........................................................................................................54

vii

LIST OF TABLES

1 Results of analysis from Experiment 1 ............................................................................15

2 Results of analysis from Experiment 2 ............................................................................25

viii

LIST OF FIGURES

1 Example picture from the picture description task ............................................................9

2 Proportion of trials on which participants matched the sentence structure of the confederate as a function of conception risk. ...................................................................26

3 Proportion of trials on which participants matched the sentence structure of the confederate as a function of the participants’ ratings of the confederates’ flirtatiousness. ....................................................................................................................27

4 Proportion of trials on which participants matched the sentence structure of the confederate as a function of the participants’ ratings of the confederates’ anger. ............28

5 Proportion of trials on which participants matched the sentence structure of the confederate as a function of participants’ level of conformity. ........................................29

6 Proportion of trials on which participants matched the sentence structure of the confederate as a function of participants’ relationship status. ..........................................31

ix

ABSTRACT

The current study uses two experiments to explore how a range of social and cognitive variables affect linguistic alignment. Experiment 1 examines the effects of a number of individual difference variables (inhibitory control, pattern learning, creativity, trait conformity, and the need to belong), while Experiment 2 focuses on one specific social variable—female fertility. In order to look at linguistic alignment (or a lack thereof), the study employs a picture description task in which dyads (each dyad is comprised of one participant and one confederate) take turns describing pictures to each other. Both experiments measured the degree to which participants aligned with confederates on sentence structure, and Experiment 2 also measured alignment of vocal pitch within the dyads. The results of Experiment 1 suggest that individual differences do not affect the amount of alignment of sentence structure. However, the results of Experiment 2 reveal that the fertility level of a female conversational partner affects the degree to which men align with women on sentence structure. Specifically, men were less likely to align with the sentence structure of a woman who was at high fertility than with the sentence structure of a woman who was at a lower level of fertility. Experiment 2 also demonstrates that men and women in conversation do align on vocal pitch. Possible explanations and implications are discussed.

1

CHAPTER 1

INTRODUCTION

When people interact, they align with each other on a number of levels. For example, previous research has shown that we align with or mimic the facial expressions (Blairy, Herrera, & Hess, 1999), postures (Berger & Hadley, 1975; Bernierie, 1988; La France, 1979, 1982; La France & Broadbent, 1976), gestures (Bavelas et al., 1988), and other behaviors (e.g., foot shaking or face rubbing, Chartrand & Bargh, 1999) of others. Additionally, people in conversation align linguistically with those they are talking to in terms of accents (Giles & Powesland, 1975), tone of voice (Neumann & Strack, 2000), rate of speech (Webb, 1969, 1972), speech rhythms (Cappella & Panalp, 1981), lexical choices (Brennan & Clark, 1996), and sentence structure (Bock, 1986; Branigan, Pickering, & Cleland, 2000). Recent research suggests that the occurrence of these types of alignment may be useful and beneficial in many ways. In fact Lakin, Valerie, Jefferis, Cheng, and Chartrand claim that behavioral alignment or mimicry “binds people together and creates harmonious relationships” (2003). In accord with this, Pickering and Garrod (2004) argue that alignment of sentence structure may have important social functions too. For example, alignment may be important for affiliation (Lakin et al., 2003, Lakin & Chartrand, 2003; Shepherd et al., 2001, Giles, Coupland, & Coupland,, 1991). Support for this idea comes from the fact that mutual liking is associated with greater similarity between people (Tenney, Turkheimer, & Oltmanns, 2009), and that alignment between people may lead to perceptions of increased similarity. Mimicry not only increases liking, but liking increases mimicry (Bernieri, 1988; Bernieri, Davis, Rosenthal, & Knee, 1994; Chartrand & Bargh, 1999; La France & Ickes, 1981; Lakin & Chartrand, 2003; Stel, Blascovich, McCall, & Vonk, 2005; Chartrand & Jefferis, 2003). Additionally, people are more likely to (behaviorally) mimic in-group

2

members than out-group members (Yabar, Johnston, Milles, & Peace, 2006), and are less likely to mimic those with stigmas than those without stigmas (Johnston, 2002). Furthermore, higher rates of alignment and mimicry occur in situations in which people have been made to feel distinct from their peer group (Uldall, Hall, & Chartrand, 2003), and among people in collectivist cultures compared to those in individualist cultures (van Baaren, Maddux, Chartrand, de Bouter, & van Knippenberg, 2003). Alignment may also aid in social situations by helping to repair social damage (Balcetis & Dale, 2005). Balcetis and Dale (2005) found that people were more likely to align (on sentence structure) with a (confederate) conversational partner who acted mean and insulting than with a partner who acted nice. Therefore, it is not surprising that many social and evolutionary psychologists believe, “our ancestors lived in an environment in which social isolates did not survive and reproduce (e.g., Buss & Kendrick, 1998; Johnson & Edgar, 1996)…failure to facilitate positive interactions, using mechanisms such as mimicry, may have led to social isolation and hence an evolutionary disadvantage” (Yabar, Johnston, Miles, & Peace, 2006, 98; Caporael, 1997, 2001; Lewin, 1943; Poirier & McKee, 1999). Despite the extant studies of mimicry and alignment, however, there are still a plethora of unanswered questions about behavioral and linguistic alignment. In the current study, I address some of these questions by examining how certain social and cognitive variables affect the degree to which people align linguistically. The paradigm I used to study linguistic alignment is structural priming—the tendency to repeat the sentence structure across utterances (e.g., Bock, 1986; Coyle & Kaschak, 2008). For example, if one hears a sentence employing the double object (DO) construction, such as “The students showed their teacher a picture,” one is more likely to use the DO construction (as in “The girl tossed the boy a ball”) rather than an alternative construction (i.e., the prepositional object—PO—construction, such as “The girl tossed a ball to the boy”) on subsequent productions. There are several reasons I believe structural priming is a good paradigm to use for studying linguistic alignment: First, structural priming occurs below conscious awareness. People are usually not aware of how they structure their sentences, do not consciously choose to align their sentence structures with conversational partners, and

3

often do not notice when structural priming occurs (Bock & Griffin, 2000; Pickering & Garrod, 2004). Second, structural priming is very pervasive. It occurs in both laboratory experiments (e.g., Branigan et al., 2000) and naturally occurring speech (e.g., Gries, 2005), from comprehension to production (e.g., Branigan et al., 2000), both within and between people (e.g., Potter & Lombardi, 1998; Branigan, et al., 2000), in both children (Savage, Lieven, Theakston, & Tomasello, 2003) and adults (e.g., Bock, 1986), in both spoken (e.g., Bock, 1986) and written (e.g., Pickering & Branigan, 1998) language, in both picture description (e.g., Coyle & Kaschak, 2008) and stem completion (e.g., Kaschak, Kutta, & Schatschneider, 2011) tasks, it transfers across different tasks (e.g., Kaschak, 2007), it is found in many different languages (e.g., English, Dutch, Japanese, and Spanish; e.g., Bock, 1986; Hartsuiker & Westenberg, 2000; Hartsuiker, Pickering, & Veltkamp, 2004), across languages within the same speaker (e.g., Loebell & Bock, 2003; Hartsuiker, Pickering & Veltkamp, 2004), among people with Broca’s aphasia (Hartsuiker & Kolk, 1998) and amnesia (Ferreira, Bock, Wilson, & Cohen, 2005), and can persist for long periods of time (e.g., Kaschak, Kutta, & Schatschneider, 2011). For these reasons, I believe structural priming is an ecologically valid paradigm ideal for examining linguistic alignment. Beyond this, researchers have suggested that structural priming plays a role in language acquisition (e.g., Savage et al., 2003) and language change (e.g., Chang, 2008). The tendency to repeat recently encountered structures can strengthen the representation of a given sentence structure (furthering the acquisition process; Tomasello, 2006), and over time may work to alter the probability of using a given structure (contributing to language change; Chang, 2008). Furthermore, very few studies have examined how social variables affect structural priming, and few studies of alignment have examined the effects of cognitive variables. Therefore, in the current study, I conducted two experiments designed to assess how a wide range of social and cognitive variables affect linguistic alignment. To do this, I had participants engage in a picture matching task developed by Branigan, et al. (2000) with a (confederate) partner. Importantly, engaging in the picture matching task gives participants the opportunity to linguistically align (or not) with another person. Additionally, I used a number of measures to examine other social and cognitive

4

variables, such as inhibitory control, conformity, and female fertility. The current experiments are discussed below in more detail.

5

CHAPTER 2

EXPERIMENT 1

Although many researchers have examined how individual differences affect explicit processes, most researchers have ignored the effects individual differences have on implicit processes such as structural priming, or assume that implicit processes are only minimally affected by individual differences (e.g., Reber, 1993; Stanovich, 2009). However, recent research by Kaufman et al. (2010) demonstrates that individual differences (e.g., differences in verbal reasoning, processing speed, and self-reported levels of openness) do in fact affect implicit learning. This research suggests that although structural priming is largely implicit (Chang, Dell, & Bock, 2006), individual differences may play an important role in the degree to which participants align with the sentence structure of a conversational partner. Because the role of individual differences in linguistic alignment has not yet been examined, it is important to do so. Thus, in Experiment 1, I investigated if and how a range of social and cognitive variables affect alignment of sentence structure. To do this, I tested whether individual differences in inhibitory control, pattern learning, creativity, trait conformity, and the need to belong affect structural priming. These variables and their possible relations to structural priming are explained below. (The specific measures used to assess these individual difference variables are discussed in the Materials.) 2.1 Inhibitory Control

Inhibitory control involves ignoring or blocking certain information from one’s attention while still processing other information. A choice to not repeat or align on sentence structure (i.e., a lack of structural priming) might require inhibiting the previous sentence structure—such as ignoring the sentence construction used by one’s

6

conversational partner—simply because processing (or comprehending) an utterance typically leads one to repeat or align on sentence structure (i.e., structural priming; Bock, 1986). So if individual differences in inhibitory control relate to structural priming, I would expect this relationship to be negative. That is, people who align more (i.e., those who demonstrate more structural priming) would be expected to be worse at inhibitory control than people who align less.

2.2 Pattern Learning

Pattern learning, at least the type of pattern learning addressed in the current study, involves implicit probability learning (Conway, Bauernschmidt, Huang, & Pisoni, 2010). (In the pattern learning task I use, participants must learn what sequences of stimuli are probable in order to succeed in the task.) Structural priming also reflects implicit learning (Chang et al., 2006), and involves implicitly detecting and tracking probabilistic information (e.g., of structural and lexical content). Thus if individual differences in pattern learning affect structural priming, I would expect people who are better at pattern learning to demonstrate higher rates of structural priming.

2.3 Creativity

Creativity requires being unique or different from others—which may involve adjusting one’s behavior (and/or thoughts) to mismatch that of others. If individual differences affect structural priming, creativity would most likely have a negative relationship with structural priming—highly creative people would be expected to exhibit less structural priming. This would be expected because structural priming involves aligning with and repeating the sentence structure of a conversational partner—that is, reusing or mimicking the sentence structure of another rather conveying the message with a different (and perhaps creative or unique) construction.

2.4 Conformity & the Need to Belong

Conformity is basically the opposite of creativity. Conformity involves adjusting one’s behavior (and/or thoughts) to go along with or match that of others. The need to belong is related to conformity in that being motivated to fulfill the need to

7

belong might lead to conformity. The need to belong is the need to feel socially connected, attached to, and accepted by others (Baumeister & Leary, 1995). Because structural priming (and other types of alignment) helps us form social bonds and functions to increase affiliation (e.g., Lakin et al., 2003), it would make sense for people who are more conforming and for those who have a stronger need to belong to exhibit more structural priming. Investigating the effects of individual differences on social and cognitive variables, such as those listed above, is important because doing so may provide insight into the nature of psychological mechanisms that underlie linguistic alignment as well as the representations that underlie comprehension and production of language (Bock, Dell, Chang & Onishi, 2007; Pickering & Ferreira, 2008). Furthermore, studying how individual differences in social and cognitive variables affect structural priming is also important because doing so can help us understand other types of alignment—which may reflect the general process of learning and may serve important roles in communication and other types of social interaction (Pickering & Ferreira, 2008). Additionally, very few studies have examined how social variables affect structural priming, and few studies of alignment have examined the effects of cognitive variables. I entertained two alternative hypotheses for the outcome of Experiment 1. One possibility is that linguistic alignment (i.e., structural priming) reflects implicit learning that is not affected by individual differences in abilities such as pattern learning and inhibitory control or personality factors such as creativity, trait conformity, and the need to belonging. If this is the case, I would expect individual differences (such as those listed above) to be irrelevant to the amount of structural priming that occurs. This finding would be consistent with traditional views that although individual differences often play a large role in explicit and conscious processes, implicit processes are much less affected (if at all) by individual differences (Reber, 1993; Stanovich, 2009). However, an alternative possibility is that individual differences do play an important role in the degree of structural priming that occurs. If individual differences do influence structural priming rates, I would expect higher pattern learning scores, higher belongingness scores, higher levels of trait conformity, lower creativity scores, and lower inhibitory control scores to be associated with increases in structural priming. These findings would lend

8

further support to Kaufman et al.’s (2010) claim that even implicit tasks can be affected by individual differences. 2.5 Method

2.5.1 Participants The participants were 89 undergraduates (67 females, 20 males, and 2 participants who did not indicate their sex) from Florida State University. Participants received partial course credit in exchange for their participation.

2.5.2 Confederates & Experimenters

Three undergraduate women served as confederates and seven undergraduate women served as experimenters.

2.5.3 Materials

Using pictures from Bock (1986), two sets of pictures (“description sets;” one for the confederate and one for the participant) were constructed. Each set consisted of 17 pictures—8 critical pictures and 9 filler pictures. Each picture had a verb typed above it, which was to be used in generating a sentence to describe the picture (see Figure 1). The critical pictures in Set 1 were the prime items. They depicted a scene that was designed to elicit a DO or PO construction, and the confederates were scripted to describe each critical item with a specific utterance (the prime sentence; see Figure 1). Critical pictures in Set 2 were also pictures designed to elicit the DO or PO construction, and provided the participant with the opportunity to use either construction (the “target” items). (Confederates always used Set 1 to produce picture descriptions, while participants always used Set 2 to produce the descriptions.) Filler pictures were designed to elicit descriptions that were neither DO nor PO constructions. Confederates produced the same set of picture descriptions for every participant. (Scripted descriptions were typed at the bottom of each picture in Set 1 to ensure that each participant was exposed to the exact same prime sentences.) Half of the primes used the DO construction, and half used the PO construction. Half of the prime-target sequences employed the same verb, and half employed different verbs between trials. The pictures in both description sets

9

were put into a fixed order, such that each critical (prime) picture in Set 1 was immediately followed in the trial sequence with a critical (target) picture from Set 2. The prime and verb manipulations were organized across trials using the following pattern: DO same verb, PO different verb, DO different verb, PO same verb. A duplicate of each description set was created to be used for identifying the picture one’s partner had just described (the “matching sets”). Matching sets were shuffled before each use.

The nurse gave the patient some water. Figure 1: Example picture from the picture description task. Note: Participants and confederates saw the picture and the verb printed on top of the page. Only confederates saw the scripted description at the bottom of the page.

Audacity sound recording software (GNU General Public License, 2006) was used to record the participants’ and confederates’ picture descriptions. Eprime (Psychology Software Tools, Inc., 2001) was used to administer several tasks designed to assess the individual differences mentioned above. Inhibitory control

10

was measured using the Simon effect task (Simon, 1969) and the flanker task (Eriksen & Eriksen, 1974; Eriksen & Schultz, 1979). In each trial of the Simon effect task, a red or blue square appears on the right or left side of the computer screen. Participants are instructed to respond to the color of the square (and meanwhile ignore its location) by pressing one key when they see a red square and a different key when they see a blue square. Importantly, one response key is on the left side of the keyboard while the other is on the right side of the keyboard. That is, the locations of the response keys correspond to the location of the stimulus. Generally, participants are faster at responding correctly when the location of the response is consistent with the location of the stimulus. However, participants who are better at inhibitory control are better at ignoring the location information (and thus focusing solely on the color of the stimulus), and therefore are generally just as quick or nearly as quick in responding in stimulus- response inconsistent trials as in stimulus-response consistent trials (Simon, 1969). Both accuracy and response time differences between consistent and inconsistent trials provide an index of individual differences in inhibitory control. In the flanker task which also measures inhibitory control (Eriksen & Eriksen, 1974; Eriksen & Shultz, 1979), participants are instructed to respond to the direction of a visual stimulus in the center of the computer screen (the “target;” i.e., they press one response key when they see a “<,” and a different response key when they see a “>”). During critical trials in this task, distracting stimuli appear on both sides of the target (and participants are instructed to ignore the distracting stimuli). On some critical trials, distracting stimuli are facing the same direction as target (as in “> > > > > > >”), while on other critical trials, distracting stimuli are facing the opposite direction as the target (as in “< < < > < < <”). Participants are typically faster at responding correctly when distracting stimuli are facing the same direction as the target than when they are facing the opposite direction as the target. However, this difference in response time is not as evident in people who are better at inhibitory control (Eriksen & Eriksen, 1974; Eriksen & Shultz, 1979). As in the Simon effect task, both accuracy and response time differences between consistent and inconsistent trials of this task provide an index of individual differences in inhibitory control.

11

In the Simon game (Baer & Horrison, 1978), which measures pattern learning abilities (Conway et al., 2010), participants see different colored squares light up one-at- a-time. (There are four squares arranged in two rows of two in the center of the computer screen.) Participants are instructed to watch the squares and pay attention to and remember the order in which they light up. When the lights stop, participants are asked to indicate the pattern of lights they saw. (There are four response keys and the location of each response key corresponds to the location of one of the stimulus squares, so participants basically use these keys to replicate the pattern they saw.) Importantly, the sequence of the lights is probabilistic, and some trials are more probabilistic (“grammatical trials”), than others (“ungrammatical trials”). Participants who are better at pattern learning (or probability learning) typically perform better on grammatical trials than on ungrammatical trials, and this discrepancy is less apparent in people who are worse at pattern learning. The difference in performance between grammatical and ungrammatical trials provides and index of one’s pattern learning ability (Conway et al., 2010). Creativity was assessed using a task developed by Mednick (1962) called the Remote Associates Test (RAT). In each trial of this task, participants are presented with three words (“problem items”) that are remotely-associated (three remotely-associated words = one “problem”). That is, all three problem items are related to each other through one other word (the “solution”). For example, participants may be presented with “sleeping/bean/trash.” (In this example, the solution is “bag,” as in sleeping bag, bean bag, and trash bag.) Generally, participants are limited in the amount of time they are given for trying to generate the solution to each problem, though any number of problems and a number of different time limits can be used. In the version of the RAT that I used, participants were given 7 seconds to come up with the solution to each of 20 problems items (Bowman & Jung-Beeman, 2003). (The problem items I used were developed by Bowman and Jung-Beeman (2003). In these problems, each problem item is related to the solution in that it can be combined with the solution to form a compound word, as in the example above.) Performance on the RAT has been shown to correlate reliably with performance on classic insight problems, and the RAT is often used to study creativity in problem solving (Bowden & Beeman, 1998; Schooler & Melcher, 1995).

12

Thus, RAT scores are believed to reflect creativity—in that participants who are more creative are typically better at identifying remote associations (and thus, solve more RAT problems) (Bowman & Jung-Beeman, 2003). Trait conformity was measured using the Conformity Scale (Mehrabian & Stefl, 1995). The Conformity Scale consists of 11 items intended to assess general tendencies toward conforming with others (e.g., “I often rely on, and act upon, the advice of others” and “Basically, my friends are the ones who decide what we do together.”). Responses are made using a scale ranging from -4 (very strong disagreement) to +4 (very strong agreement). Higher scores on the Conformity Scale indicate higher levels of trait conformity (Mehrabian & Stefl, 1995). I used the Belongingness Scale (Leary, Kelly, Cottrell, & Schreindorfer, 2005) to measure belongingness. The Belongingness Scale consists of 10 items intended to assess one’s general need to belong (e.g., “I try hard not to do things that will make other people avoid or reject me” and “It bothers me a great deal when I am not included in other people’s plans”). Responses are made using a scale ranging from 1 (strongly disagree) to 5 (strongly agree). Higher scores on the Belongingness Scale indicate higher levels of belongingness (Leary et al., 2005). A questionnaire packet was used to assess participants’ demographic information (e.g., sex and age). 2.5.4 Procedure

At the start of each session, an experimenter arrived at a participant waiting room and said that she was looking for “[participant name] and [confederate name].” If the participant was present, the experimenter guided him/her to a lab room and asked him/her to sit at a desk with two chairs side-by-side. During this time, the confederate was in a separate room and out-of-sight. The experimenter told the participant she was going to check the waiting room again to see if the other participant had shown up, and would return shortly. The experimenter closed the door and left, returning with the confederate a few minutes later. The experimenter asked the confederate to sit in the open chair. After having the participant and confederate fill out consent forms, the experimenter explained that the purpose of the study was to examine how people work together, and

13

that the two of them would work together to complete a picture matching task. The experimenter asked the participant and confederate to move to the other side of the room and sit at desks separated by a divider. The experimenter then explained the picture matching task (Branigan et al., 2000). The pair was told that they would take turns describing pictures to each other, and that they would have two goals—to create a sentence to describe each picture, and to find the picture that matches the description given by their partner. The participant and confederate were each given two stacks of pictures—a “description stack” for them to describe pictures to their partner and a “matching stack” for them to use to find the pictures described to them. The confederate always went first in this task. The confederate was asked to begin by creating a sentence describing the first picture in her description stack, using the verb written at the top of the picture. She was told to say this sentence aloud. The participant was told that he/she was to find the picture his/her partner described in his/her matching stack. The participant was then instructed to create a sentence for the first picture in his/her description stack, using the specified verb. The confederate was instructed to find the picture in her matching stack. The pair was instructed to keep taking turns until all pictures had been described and matched. The experimenter asked if the participant or confederate had any questions, reminded them to listen carefully to what their partner said so that they could find the right picture, and started the audio recorder. The task was completed as described, except that the confederate used a script to describe each picture. After the picture matching task, participants (individually) completed several tasks on a computer (the tasks used as individual difference measures, described above— both the instructions for the tasks and tasks themselves were administered via a computer) and filled out the questionnaire packet. Participants completed these tasks in a fixed order: they did the Simon effect task first, then they did the RAT task, next they did the flanker task, then they completed the questionnaire packet, the Conformity Scale, and the Belongingness Scale, and finally they did the Simon game last. Upon completion of the Simon game, participants were debriefed. 2.5.5 Design & Analysis

Audio files from the picture match task were transcribed, and the participants’

14

responses to the critical pictures were scored as DO, PO, or “other,” following the criteria described in Kaschak (2007) by research assistants blind to the hypotheses of the current study. Trials on which an “other” response was made (5% of trials) were excluded from subsequent analyses. The dependent measure in this experiment was Match, which coded whether participants used the same sentence structure to describe the target picture as the confederate used in the directly preceding prime sentence (i.e., whether participants showed structural priming on a given trial). A trial was coded as “1” if the participant matched constructions (e.g., the participant produced a DO target description after hearing a DO prime sentence), and a “0” if the participant produced the alternative construction (e.g., producing a PO target following a DO prime sentence). Mixed logit model analysis of the target descriptions was performed to predict the logit-transformed likelihood of a target using the same syntactic construction as the confederate’s prime sentence. I performed the analysis with Match as the dependent measure, and participants and items as crossed random factors. The following variables were also included as predictors: Conformity Score, Belongingness Score, Creativity (RAT score), Inhibitory control A (response time and accuracy in the Simon effect task), Inhibitory control B (response time and accuracy in the flanker task), Pattern learning (score in the Simon game), Prime type (DO vs. PO), Verb repetition (Same verb vs. Different verb on prime and target trials), and the Prime type x Verb repetition interaction. (Note that the pattern learning data was entered into a separate analysis from that of the other individual difference variables to avoid problems with missing data due to computer error.) To avoid issues with collinearity, all variables were grand-mean centered before being entered into the analyses. The regression analyses were performed using the HLM statistical package (Raudenbush, Bryk, & Congdon, 2004).

Full document contains 64 pages
Abstract: The current study uses two experiments to explore how a range of social and cognitive variables affect linguistic alignment. Experiment 1 examines the effects of a number of individual difference variables (inhibitory control, pattern learning, creativity, trait conformity, and the need to belong), while Experiment 2 focuses on one specific social variable--female fertility. In order to look at linguistic alignment (or a lack thereof), the study employs a picture description task in which dyads (each dyad is comprised of one participant and one confederate) take turns describing pictures to each other. Both experiments measured the degree to which participants aligned with confederates on sentence structure, and Experiment 2 also measured alignment of vocal pitch within the dyads. The results of Experiment 1 suggest that individual differences do not affect the amount of alignment of sentence structure. However, the results of Experiment 2 reveal that the fertility level of a female conversational partner affects the degree to which men align with women on sentence structure. Specifically, men were less likely to align with the sentence structure of a woman who was at high fertility than with the sentence structure of a woman who was at a lower level of fertility. Experiment 2 also demonstrates that men and women in conversation do align on vocal pitch. Possible explanations and implications are discussed.