Evidence for Human-Like Conversational Strategies in An

EVIDENCE FOR HUMAN-LIKE CONVERSATIONAL STRATEGIES IN AN

ERIN NATANNIE COLBERT-WHITE

(Under the Direction of Dorothy Fragaszy)

ABSTRACT

Researchers have established many similarities in the structure and function of human and avian communication systems. This dissertation investigated one unique nonhuman communication system—that of a speech-using African Grey parrot. In the same way that humans learn communicative competence (i.e., knowing what to say and how given a particular social context), I hypothesized that Cosmo the parrot‘s vocalizations to her caregiver, BJ, would show evidence of similar learning. The first study assessed turn-taking and the thematic nature of Cosmo‘s conversations with BJ.

Results confirmed that Cosmo took turns during conversations which were very similar to humans‘ average turn-taking time. She also maintained thematically linked dialogues, indicating strategic use of her vocal units. The second study investigated Cosmo‘s ability to take BJ‘s auditory perspective by manipulating the distance between the two speakers.

As predicted, Cosmo vocalized significantly more loudly when her owner was out of the room, and those vocalizations classified as social (e.g., kiss sounds) were uttered significantly more loudly than vocalizations which were considered nonsocial (e.g., answering machine beep sounds). These results suggested Cosmo may be able to take the perspective of a social partner, an ability others have documented in Greys using alternative tasks. Strategic use of vocalizations was revisited in Studies 3 and 4. The third study examined Cosmo‘s requesting behavior by comparing three separate corpora

(i.e., bodies of text): Cosmo‘s normal vocalizations to BJ without requesting, Cosmo‘s vocalizations following a denied request, and Cosmo‘s vocalizations following an ignored request. Cosmo treated all three speech contexts differently; distributions and rate of use of speech, nonspeech, social, and nonsocial vocalizations differed significantly. The results of the fourth study supported the third study‘s findings. While

Cosmo did not use phrases which BJ classified as ―jokes‖ significantly more often following a refused request (which would indicate strategic use of vocalizations to manipulate affect), her vocalization patterns indicated that she perceived the refusal condition as being different from normal conversation. While the results can be compared to child sociocognitive development, my findings may also be extended to wild

Greys‘ communication system to learn more about the species‘ natural capabilities.

INDEX WORDS: African Grey parrot, language, communication, turn-taking,

communicative competence, perspective-taking, requesting, humor

EVIDENCE FOR HUMAN-LIKE CONVERSATIONAL STRATEGIES IN AN

AFRICAN GREY PARROT'S SPEECH

ERIN NATANNIE COLBERT-WHITE

BS, Denison University, 2007

MS, University of Georgia, 2009

A Dissertation Submitted to the Graduate Faculty of The University of Georgia in Partial

Fulfillment of the Requirements for the Degree

DOCTOR OF PHILOSOPHY

ATHENS, GEORGIA

2013

Erin Natannie Colbert-White

EVIDENCE FOR HUMAN-LIKE CONVERSATIONAL STRATEGIES IN AN

AFRICAN GREY PARROT'S SPEECH

ERIN NATANNIE COLBERT-WHITE

Major Professor: Dorothy Fragaszy Committee: Gary Baker Michael Covington Adam Goodie

Electronic Version Approved:

Maureen Grasso Dean of the Graduate School The University of Georgia May 2013

DEDICATION

I dedicate this dissertation to Betty Jean Craige and Cosmo. Without them, my graduate career would have been far less exciting.

ACKNOWLEDGEMENTS

To my adviser Dr. Dorothy Fragaszy, I say thank you for your patience and support. To committee members Drs. Michael Covington, Adam Goodie and Gary

Baker, I appreciate your thoughtful input throughout the process, and great conversations along the way. I thank my family, friends, and mentors who supported me throughout the years. In addition, I offer special thanks to Hannah Hall, Marisol Macias, and Rachel

Callery for their help with data collection and organization, as well as Dominic Byrd for his artistic ability and encouragement. Thank you to Michael Amlung, Sarah Whitaker, and (especially!) Vijay Veeraghattam for second-semester writing motivation. I also thank the University of Georgia State Botanical Gardens for existing. The gardens helped me get over my brief battle with senioritis in January 2013. I learned that giant banana leaves and papaya trees are an inspirational and motivational force. Above all, I thank both Betty Jean Craige and Cosmo for their participation and excitement, along with Betty Jean‘s continued support and confidence in me to show the world just how phenomenal Cosmo is!

TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS ...... v

LIST OF TABLES ...... viii

LIST OF FIGURES ...... x

CHAPTER

1 INTRODUCTION AND LITERATURE REVIEW ...... 1

LANGUAGE IN ACTION ...... 1

NONHUMAN LANGUAGE RESEARCH ...... 3

2 STUDY #1: TURN-TAKING AND DIALOGUE ...... 14

INTRODUCTION ...... 16

METHOD ...... 19

RESULTS ...... 22

DISCUSSION ...... 26

3 STUDY #2: PERCEPTUAL PERSPECTIVE-TAKING ...... 57

INTRODUCTION ...... 59

METHOD ...... 64

RESULTS ...... 67

DISCUSSION ...... 71

4 STUDY #3: DENIED AND IGNORED REQUESTS ...... 91

INTRODUCTION ...... 93

METHOD ...... 98

RESULTS ...... 101

DISCUSSION ...... 106

5 STUDY #4: MANIPULATING AFFECT THROUGH ―JOKES‖ ...... 136

INTRODUCTION ...... 138

METHOD ...... 141

RESULTS ...... 145

DISCUSSION ...... 147

6 GENERAL CONCLUSIONS ...... 170

SUMMARY OF FINDINGS FROM PRESENT STUDY ...... 171

THE ROLE OF ECOLOGICAL RELEVANCE ...... 176

FUTURE DIRECTIONS ...... 179

CONCLUSIONS...... 180

vii

LIST OF TABLES

Page

Table 2.1: Nonword sound coding scheme ...... 36

Table 2.2: Length, turn, interruption, and theme data for each conversation ...... 38

Table 2.3: Conversation initiations made by Betty Jean...... 48

Table 2.4: Single and multiple turns within turns ...... 49

Table 2.5: Longest single-turn turn-taking event ...... 50

Table 2.6: Common themes in dialogues and their respective frequencies ...... 51

Table 3.1: Nonword sound coding scheme ...... 83

Table 3.2: Social and nonsocial vocalization classification ...... 85

Table 3.3: Volume (dB) of vocalizations for classifications and social contexts ...... 88

Table 4.1: Nonword sound coding scheme ...... 117

Table 4.2: Social and nonsocial vocalization classification ...... 119

Table 4.3: Twenty most frequently uttered vocalizations and respective frequencies ....122

Table 4.4: Conversation excerpts containing repetition and redirection ...... 123

Table 4.5: Conversation excerpts containing interruption ...... 125

Table 4.6: Conversation except containing a denied then met request ...... 127

Table 4.7: Conversation excerpts containing persistence and refocusing ...... 129

Table 4.8: Conversation excerpt containing ―be a good bird‖ and ―Cosmo don‘t bite

okay‖ ...... 131

Table 5.1: Nonword sound coding scheme ...... 157

viii

Table 5.2: List of Cosmo‘s ―jokes‖ used in analyses...... 159

Table 5.3: Example data preparation for multiple requests ...... 160

Table 5.4: Distribution of jokes across conflict events for each taped session ...... 161

Table 5.5: Twenty most frequently uttered vocalizations and associated frequencies ....163

Table 5.6: Joke use frequencies ...... 165

Table 5.7: Outlier requesting event from Session 4 ...... 166

LIST OF FIGURES

Page

Figure 2.1: Illustration of the testing space ...... 52

Figure 2.2: Frequency plot for Cosmo‘s turn take-up time ...... 53

Figure 2.3: Frequency plot for BJ‘s turn take-up time...... 54

Figure 2.4: Frequency plot for conversation length ...... 55

Figure 2.5: Frequency plot for coupled turns...... 56

Figure 3.1: Illustration of the testing space ...... 89

Figure 3.2: Social context and vocalization classification data ...... 90

Figure 4.1: Illustration of the testing space ...... 133

Figure 4.2: Nonrandom distributions of speech and nonword vocalizations...... 134

Figure 4.3: Nonrandom distributions of social and nonsocial vocalizations ...... 135

Figure 5.1: Illustration of the testing space ...... 168

Figure 5.2: Total number of ―jokes‖ vocalized during requesting events ...... 169

CHAPTER 1

INTRODUCTION AND LITERATURE REVIEW

Language in Action

Language can both transfer information and incite action in others. Regardless of the function, instances of language use are called speech acts. A speech act refers to any interaction involving information transfer from sender to receiver. Austin (1962) explained that all utterances have locutionary (e.g., literal meaning), illocutionary (e.g., speaker‘s intention), and perlocutionary (e.g., effect on listener) properties. In this way, the intention behind a speech act can be literal (as in ―It‘s cold in here‖ literally translating to the temperature in a room) or not literal (as in ―It‘s cold in here‖ being intended to solicit the listener to close the window).

Language can be studied from two angles: structural and functional (e.g., de

Saussure, 1916/1983; Premack, 1986). de Saussure (1916/1983) conceptualized language langue (language) and parole (speaking). Langue refers to the structural features of language. That is to say, those who study structural features are concerned with the rules, syntax, and grammar of a particular language. The parole de Saussure described pertains to the functional properties of language. These properties reflect the way language is used in action to accomplish goals, and are those which exist in conversation with others, in written texts, and even during spoken language while alone (Paltridge, 2006).

Examples of functional features include joint-attention to the conversation and/or object

being discussed, turn-taking, eye contact, body posture, volume, intonation, and pitch.

All of these features provide valuable socially-relevant content which is not expressed directly by the words or signs being used in conjunction with them. These ―social dynamics‖ and behaviors associated with language are undeniably important, and without them, accurate transfer of information would be impossible. The studies comprising this dissertation explore four functional strategies of language—that is to say, language in action.

Many have highlighted the relevance of the functional features of language (e.g.,

Austin, 1962; Searle, 1969; de Saussure, 1916/1983; Tomasello, 2003). Rather than focusing on the structural features, these individuals described language as going beyond the grammar and literal meaning of words. To them, language involved ‗doing things‘ to achieve goals. Frequent social interaction with others provides children with the opportunity to learn to engage strategies to act with language. Through extensive communicative experience with others, children develop the ability to employ functional strategies and use them to their advantage during social interaction (Slobin, 1985).

Slobin and others (e.g., Tomasello, 2003) assert that many of these strategies exist far before to true language acquisition, and as early as 1 year. Joint-attention, turn-taking, eye-contact, body language, modifying volume and pitch, requesting behavior, interruption, and humor are all learned strategies that can enhance or modify a language- learning child‘s goal-oriented speech. The question to be answered in this dissertation was whether a home-reared nonhuman speaker, through extensive social interaction, could also learn to engage such strategies.

Nonhuman Animal Language Research

Over the past 50 years, pioneering work with apes, marine mammals, and avian species has defined the nonhuman animal (hereafter NHA) language subfield. Attention was first devoted to nonhuman primates during the 1950s and continues into the present.

Due to their inability to articulate speech sounds (Hayes & Hayes, 1951), apes are predominantly taught to use gestural signing or keyboards with referent symbols.

Lexigram work with Kanzi (Pan paniscus, e.g., Savage-Rumbaugh, 1986; Savage-

Rumbaugh et al., 1986) and Lana (Pan troglodytes, e.g., Rumbaugh, 1977; Beran, Pate,

Richardson, & Rumbaugh, 2000), as well as American Sign Language training with

Koko (Gorilla gorilla, e.g., Patterson & Cohn, 1990), are a few examples of projects that provided insights into the structural language training process with nonhuman primates.

Many have noted the similarities between child language acquisition and lexigram or sign acquisition in apes like Kanzi (e.g., Bonvillian & Patterson, 1993; Savage-

Rumbaugh et al., 1993). For example, like children, Kanzi learned the meaning of symbols through observation and social interaction with his handlers. In addition, Kanzi comprehended more spoken English than was originally thought, which is true of pre- linguistic and language-learning children. For comparison, Kanzi correctly carried out

59% of novel commands (e.g., Take the telephone outdoors), while a 2-year old child in

Savage-Rumbaugh and colleagues‘ study carried out 54%.

In the 1980s, NHA language researchers began testing marine mammals on rule- based symbolic language codes. Lou Herman pioneered gestural language research with bottlenose dolphins (Tursiops truncatus, e.g., Herman, 1987; Herman et al., 1993). In one study, dolphins were presented with two types of novel commands—lexically novel,

where known words were inserted into other trained syntactic frames, or syntactically novel, where known words were reorganized into novel syntactic frames (Herman, 1987).

Two dolphins, Ake and Phoenix, performed better than chance (59-87% accuracy) on both types of novel commands. According to Herman, the dolphins‘ performance demonstrated an understanding and flexible use of the syntactic rule subject-verb-object.

Another marine mammal, the California sea lion (Zalophus californianus), has also shown evidence of rudimentary syntactical rule understanding (e.g., Gisiner &

Schusterman, 1992). Gisiner and Schusterman presented one subject, Rocky, with probe trials of unfamiliar combinations of gestured signs such as commanding actions to be done without specifying an object or modifier (e.g., color), or specifying a modifier and action but no object. According to the authors, Rocky‘s responses to these anomalous combinations indicated she had learned conditional relations (i.e., actions require objects, objects require modifiers when options are available). For example, in one trial, when presented with the gibberish command Large Fetch (modifier-action), she did not fetch any object. Results such as these with dolphins and sea lions have shown that nonprimate mammals may also be able to learn grammatical rules involving artificial languages.

Most relevant to this dissertation is Irene Pepperberg‘s work with African Grey parrots (see Pepperberg, 1999 for review). Unlike apes and marine mammals, parrots are able to use speech (rather than gestures or signs) to manipulate and use the rules of language. Pepperberg‘s (1990) training procedure, termed ‗referential mapping‘ is a three-part process involving connecting labels to objects—a series of training techniques called model/rival (M/R) training (see Pepperberg, 1981, 1988; Todt, 1975). M/R training is then followed by frequent repetition of labels during interaction with objects.

According to Pepperberg (1990), referential mapping approaches fast mapping, one process by which young children learn object labels (Carey & Bartlett, 1978). Through a variety of studies spanning three decades, one subject, Alex, demonstrated number comprehension for 0-6 (Pepperberg & Gordon, 2005); the ability to learn that objects can belong to multiple classes such as color, shape, and size (Pepperberg, 1983; Pepperberg

& Brezinsky, 1991); and he spontaneously recombined labels in the presence of novel objects, indicating an understanding that labels were built of individual units

(Pepperberg, 2007).

Despite Pepperberg‘s rich body of research with African Greys, she has not published any work on her or her trainers‘ spontaneous conversations with Alex. With the exception of one study on solitary sound-play while alone (Pepperberg, Breese, &

Harris, 1991), all of Pepperberg‘s work has focused on Greys answering questions posed to them by experimenters. Further, a majority of all NHA language research has emphasized the structural features of language. Despite this, there are some who see the importance of considering the functional features of language in NHA research. For example, Premack (1986) wrote:

Neither did the linguist help the animal language inquiry in quite the ways that were needed. The questions for which one needed answers were very abstract: What is language? On the basis of what properties do we identify a system as language? The answers could emphasize either the structural or function properties of the system, though the ideal answer would relate the two, showing the structures on which particular functions depend. (p. 29)

Even though the NHA language debate has until now largely ignored the functional features (for logistical, methodological, or other reasons), language is a harmony between the grammatical rules and the knowledge and use of social graces which are acquired through interaction. This gap in the literature is where the current work picks up—

addressing the influence of social context and social partner affect on spontaneous vocalization patterns in a highly social, highly intelligent nonhuman speaker.

Simply put, this dissertation represents a shift in the focus of NHA language research. Rather than investigating structural features of language, I argue that attention should be turned to how social animals use artificial vocal or gestural codes to accomplish goals during communicative interactions with social partners. In particular, this series of four studies explores features of a nonhuman speaker‘s conversations with a human social partner and the extent to which socially-learned functional features of language appear.

To learn more about how a nonhuman speaker utilized vocalizations during conversation, one African Grey parrot‘s utterances were analyzed for use of four pre- linguistic communicative strategies during taped conversations with her owner. The qualifier ―pre-linguistic‖ indicates simply that the speaker—regardless of species—need not have acquired ―true language‖ in order to use the strategy. For example, turn-taking is one functional strategy which appears in pre-linguistic children. The presence of any of the strategies in a nonhuman would indicate nothing about whether the speaker possessed true language as it is defined in humans. For this reason, the distinction of

―pre-linguistic‖ is made.

The exploration of conversational strategies was conducted using transcriptions of conversations between a then 6-year-old female African Grey named Cosmo and her human caregiver, Betty Jean. Both quantitative and qualitative observations were made of Cosmo‘s (and sometimes Betty Jean‘s) vocalizations. The analyses for this series of investigations were particularly exciting because unlike other NHAs and most parrots,

Cosmo possessed a lexicon of over 250 units (i.e., words, nonword sounds, and phrases), and she was able to use these words in a contextually and thematically relevant manner

(Colbert-White, Covington, & Fragaszy, 2011; Kaufman, Colbert-White, & Burgess,

2013, in press). Thus, this investigation extended my graduate work, which has involved adapting methods and applied theories associated with linguistics and child language acquisition to a nonhuman speaker.

Based on a review of the literature, I selected four pre-linguistic communication strategies for which methodological modifications for a nonhuman speaker could be made. I analyzed transcriptions of conversations between Cosmo and Betty Jean, and explored the dataset to determine: (1) whether Cosmo engages in turn-taking and thematically-linked dialogue with Betty Jean, (2) whether Cosmo engages in perceptual perspective-taking by modifying the volume of her vocalizations to Betty Jean in relation to an auditory barrier (i.e., distance), (3) whether Cosmo modifies her requesting behavior in predictable ways depending upon whether a request is denied or ignored by

Betty Jean, and (4) whether Cosmo uses vocalizations Betty Jean calls ―jokes‖ more often during times of relationship conflict (i.e., refusal of Cosmo‘s requests).

The results of this series of studies contribute to what is known regarding learned social strategies in conversations between nonhuman speakers and humans. Given the numerous parallels between parrots‘ and language-learning children‘s vocal behavior, when child development literature was available, I aligned Cosmo‘s vocal behavior with the emergence of these four strategies in children. Researchers such as Irene Pepperberg have likened African Greys‘ cognitive abilities to that of 4-6 year olds in numerical skills, object class learning, and mutual exclusivity learning (e.g., Pepperberg, 1999, 2006);

however, there is a lack of research on Greys‘ competency in the realm of social cognition. In addition to drawing parallels between Cosmo‘s behavior and children‘s behavior, findings with Cosmo regarding conversational strategies could be transferred to understand better how wild African Greys might use their species-typical repertoires.

Most importantly, however, my dissertation introduces a new perspective to the animal language field, that of the sociocognitive abilities of a nonhuman speaker.

References

Austin, J. L. (1962). How to do things with words. Cambridge, MA: Harvard

University Press.

Beran, M. J., Pate, J. L., Richardson, W. K., & Rumbaugh, D. M. (2000). A

chimpanzee‘s (Pan troglodytes) long-term retention of lexigrams. Animal

Learning & Behavior, 28, 201-207. doi:10.3758/BF03200255

Bonvillian, J. D., & Patterson, F. G. P. (1993). Early sign language acquisition in

children and gorillas: Vocabulary content and sign iconicity. First Language, 13,

315-338. doi:10.1177/014272379301303903

Carey, S., & Bartlett, E. (1978). Acquiring a single new word. Papers and Reports on

Child Language Development, 15, 17-29.

Colbert-White, E. N., Covington, M. A., & Fragaszy, D. M. (2011). Social context

influences the vocalizations of a home-raised African Grey parrot (Psittacus

erithacus erithacus). Journal of Comparative Psychology, 125, 175-184. doi:

10.1037/a0022097 de Saussure, F. (1983). Course in general linguistics. Chicago: Open Court

Publishing. (Original work published 1916)

Gisiner, R., & Schusterman, R. J. (1992). Sequence, syntax, and semantics: Responses

of a language-trained sea lion (Zalophus californianus) to novel sign

combinations. Journal of Comparative Psychology, 106(1), 78-91.

Hayes, K. J., & Hayes, C. (1951). The intellectual development of a home-raised

chimpanzee. Proceedings of the American Philosophical Society, 95(2), 105-109.

Herman, L. M. (1987). Receptive competencies of language-trained animals. In J. S.

Rosenblatt, C. Beer, M.-C. Busnel, & P. J. B. Slater (Eds.), Advances in the study

of behavior, vol. 17, (pp. 1-60). Orlando, FL: Academic Press.

Herman, L. M., Kuczaj, S. A., & Holder, M. D. (1993). Responses to anomalous

gestural sequences by a language-trained dolphin: Evidence for processing of

semantic relations and syntactic information. Journal of Experimental

Psychology, 122, 184-194. doi:10.1037/0096-3445.122.2.184

Kaufman, A. B., Colbert-White, E. N., & Burgess, C. (2013, in press). Higher-order

semantic structures in an African Grey parrot's vocalizations: Evidence from the

Hyperspace Analog to Language (HAL) model. Animal Cognition.

Paltridge, B. (2006). Discourse analysis. London: Continuum.

Patterson, F. G., & Cohn, R. N. (1990). Language acquisition by a lowland gorilla:

Koko‘s first ten years of vocabulary development. WORD, 41(3), 97-142.

Pepperberg, I. M. (1981). Functional vocalizations by an African Grey parrot (Psittacus

erithacus). Zeitschrift fürTierpsychologie, 55, 139-160. doi:10.1111/j.1439-

0310.1981.tb01265.x

Pepperberg, I. M. (1983). Cognition in the African Grey parrot: Preliminary evidence

for auditory/vocal comprehension of the class concept. Learning & Behavior, 11,

179-185. doi:10.3758/BF03199646

Pepperberg, I. M. (1988). An interactive modeling technique for acquisition of

communication skills separation of labeling and requesting in a psittacine subject.

Applied Psycholinguist, 9, 59-76.

doi:http://dx.doi.org/10.1017/S014271640000045X

Pepperberg, I. M. (1990). Referential mapping: A technique for attaching functional

significance to the innovative utterances of an African Grey parrot (Psittacus

erithacus). Applied Psycholinguistics, 11, 23-44.

doi:http://dx.doi.org/10.1017/S0142716400008274

Pepperberg, I. M. (1999). The Alex studies: Cognitive and communicative abilities of

Grey parrots. Cambridge, MA: Harvard University Press.

Pepperberg, I. M. (2006). Cognitive and communicative abilities of Grey parrots.

Applied Animal Behaviour Science, 100, 77-86.

doi:10.1016/j.appanim.2006.04.005

Pepperberg, I. M. (2007). Grey parrots do not always ‗parrot‘: The roles of imitation

and phonological awareness in the creation of new labels from existing

vocalizations. Language Sciences, 29, 1-13. doi:10.1016/j.langsci.2005.12.002

Pepperberg, I. M., Brese, K. J., & Harris, B. J. (1991). Solitary sound play during

acquisition of English vocalizations by an African Grey parrot (Psittacus

erithacus): Possible parallels with children‘s monologue speech. Applied

Psycholinguistics, 12, 151-178.

doi:http://dx.doi.org/10.1017/S0142716400009127

Pepperberg, I. M., & Brezinsky, M. V. (1991). Acquisition of a relative class concept

by an African gray parrot (Psittacus erithacus): Discriminations based on relative

size. Journal of Comparative Psychology, 105, 286-294. doi:10.1037/0735-

7036.105.3.286

Pepperberg, I. M., & Gordon, J. D. (2005). Number comprehension by a Grey parrot

(Psittacus erithacus), including a zero-like concept. Journal of Comparative

Psychology, 119, 197-209. doi:10.1037/0735-7036.119.2.197

Premack, D. (1986). “Gavagai” or the future history of the animal language

controversy. Cambridge, MA: MIT Press.

Rumbaugh, D. M. (1977). Language learning by a chimpanzee: The Lana project.

New York: Academic Press.

Savage-Rumbaugh, S. (1986). Ape language: From conditioned response to symbol.

New York: Columbia University Press.

Savage-Rumbaugh, S., McDonald, K., Sevcik, R. A., Hopkins, W. & Rubert, E. (1986).

Spontaneous symbol acquisition and communicative use by pygmy chimpanzees

(Pan paniscus). Journal of Experimental Psychology, 115, 211-235.

doi:10.1037/0096-3445.115.3.211

Savage-Rumbaugh, E. S., Murphy, J., Sevcik, R. A., Brakke, K. E., Williams, S. L., &

Rumbaugh, D. M. (1993). Language comprehension in ape and child.

Monographs of the Society for Research on Child Development, 58, Nos. 3-4.

Searle, J. R. (1969). Speech acts. London: Cambridge University Press.

Slobin, D. (1985). Cross-linguistic evidence for the language-making capacity. In D. I.

Slobin (Ed.), The crosslinguistic study of language acquisition, vol 2: Theoretical

issues, (pp. 1157-1256). Hillsdale, NJ: Lawrence Erlbaum Associates.

Todt , D. (1975). Social learning of vocal patterns and modes of their applications in

Grey parrots (Psittacus erithacus). Zeitschrift fürTierpsychologie, 39, 178-188.

doi:10.1111/j.1439-0310.1975.tb00907.x

Tomasello, M. (2003). Constructing a language: A usage-based theory of language

acquisition. Cambridge, MA: Harvard University Press.

CHAPTER 2

TURN-TAKING AND THEMATICALLY-LINKED DIALOGUE IN

CONVERSATIONS BETWEEN AN AFRICAN GREY PARROT AND HER HUMAN

CAREGIVER1

______

1 Colbert-White, E. N., & Hall, H. C., to be submitted

Abstract

For humans, some nonhuman primates, and some birds, turn-taking is an important feature of vocal interactions with social partners. In parrots, antiphonal duetting

(synchronized back-and-forth bouts of vocalizations) promotes and strengthens bonds.

Given the ecological relevance of turn-taking to wild parrots, we sought to describe the spontaneous vocal interactions between an African Grey parrot, Cosmo, and her human caregiver, Betty Jean. Analyses were made at the levels of 1) all turn-taking behavior, 2) conversations which contained ≤ 3 s between turns, and 3) dialogues which contained thematically-linked content with ≤ 3 s between turns. Cosmo‘s typical turn take-up time was within 0.5 s of the average range for humans, indicating that Cosmo most likely learned the nuanced, but very important, conversational feature of appropriate turn take- up time from Betty Jean. She also initiated and maintained thematically-linked turn- taking bouts for up to 8 back-and-forth turns with Betty Jean. While there was evidence that Cosmo uses her repertoire more for social functions rather than conversational or literal function, her use of thematically-linked utterances confirmed our hypothesis that her responses to Betty Jean are heavily dependent upon what BJ says to her. The finding of novel generation of thematically-linked vocalizations in Cosmo‘s conversations with

Betty Jean may provide an indirect look into the capabilities of wild duetting African

Greys.

Keywords: African Grey parrot, Psittacus erithacus, turn-taking, duetting conversation, language, communication

Introduction

In English conversations, it is customary and expected that one individual speaks at a time. Once a thought has been completed, the speaker may signal the end of the turn with a pause, a vocalization such as ―mmm‖ or ―anyway,‖ eye contact, body posture, or changes in intonation or volume (Paltridge, 2006). A speaker can also assume the turn or take over the turn through overlap or interruption. Regardless of how speakers are signaled, normal English conversation consists of turn-taking.

Though learned through social interaction, turn-taking is biologically linked to language acquisition as evidenced by some language-impaired children consistently violating turn-taking rules (e.g., Adams & Bishop, 1989). The learning process begins at a very early age. Infants as young as 3 months modify their vocalizations in response to turn-taking patterns. Bloom, Russell, and Wassenberg (1987) showed that when mothers interacted with their infants using structured turn-taking, the infants‘ vocalizations were more speech-like (syllabic) than nonspeech-like (vocalic). When mothers violated the turn-taking structure, the opposite was observed. As infants develop the use of words, adherence to turn-taking rules increases. By 18 months, pre-linguistic infants readily engage in babble turn-taking with caregivers, using eye-gaze to signal ends of turns (e.g.,

Rutter & Durkin, 1987). Sensitivity to pause length within dialogue develops later, around 2-3 years, at which time toddlers equate longer pauses to ―no response‖ from the social partner, and shorter pauses to an end of a turn (Craig & Gallagher, 1983).

According to Gallagher and Craig (1981), by age 4, children consistently use turn-taking in dialogue and rarely overlap other speakers. The authors report only two major categories of vocalization overlap in their dataset: accidental interruption and intentional

interruption which reflected tension or frustration. After 4 years, children‘s turn-taking time approximates the normal adult turn-taking range of ~0.8-1.5 s (Garvey & Berninger,

1981).

Turn-taking is not a communicative feature limited to human interactions, nor are age-defined differences in turn-taking rule violations. In at least one species of nonhuman primate, the Campbell‘s monkey (Cercopithecus campbelli), juveniles were found to violate turn-taking rules more often than adults during vocal interactions

(Lemasson et al., 2011). In a playback follow-up study, adults but not juveniles attended differentially to violations in turn-taking. With these results, the authors concluded turn- taking and other learned conversational rules may exist in other species‘ communication systems.

Among nonhumans, asynchronously coordinated vocalizations (i.e., antiphonal duetting) are presumed to serve a variety of functions such as joint resource defense, pair- bond formation and maintenance, mate guarding, and identity signaling (e.g., Méndez-

Cárdenas & Zimmermann, 2009; Sonnenschein & Reyer, 1983; see Douglas & Mennill,

2010, for review). Taxonomically, antiphonal duetting is found most commonly in birds and primates. Among birds, tropical and subtropical songbirds (e.g., slate-colored boubou, Laniarius funebris, Sonnenschein & Reyer, 1983; spotted palm thrush,

Cichladusa guttata, Todt & Fiebelkorn, 1980) frequently engage in antiphonal duetting.

Among nonhuman primates, some monogamous species (e.g., coppery titi monkeys,

Callicebus cupreus, Müller & Anzenberger, 2002; Milne Edwards‘ sportive lemurs, lepilemur edwardsi, Méndez-Cárdenas & Zimmermann, 2009; indris, Indri indri,

Thalmann, Geissmann, Simona, & Mutschler, 1993; see Haimoff, 1986, for review of

African and Asian primates), as well as pygmy marmosets (Cebuella pygmaea, Snowdon

& Cleveland, 2005) duet antiphonally.

Parrots, which are highly affiliative with flockmates, also engage in asynchronous vocalization to promote and strengthen bonds (e.g., Nottebohm, 1972; Serpell, 1981).

While little is known regarding African Grey parrots‘ natural history, Pepperberg (1999) provided her own unpublished observation of Greys engaging in antiphonal duetting at a nest site and posited a precopulatory or territory defense function. Antiphonal duetting between Greys and humans has also been documented in the lab (e.g., Bottoni, Massa, &

Boero, 2003), as well as in home-reared environments during times of visual separation from a caregiver (Colbert-White, Covington, & Fragaszy, 2011).

The current study explored turn-taking (i.e., antiphonal duetting) in conversations between one African Grey parrot named Cosmo and her human caregiver, Betty Jean.

While previous studies explored possible functions of turn-taking (e.g., maintaining contact, pair bonding, courtship, territory defense), the current study sought to describe features of the conversations and compare them to human conversation. Analyses included assessments of turn length, turn take-over time, cohesion of the content from turn to turn, as well as the context surrounding instances of interruption.

In order to make comparisons to human conversation, an accepted human definition of turn-taking was used. This was back and forth vocal interaction which excluded interruption or simultaneous talking (e.g., Schegloff, 1968). To this definition, we added that the elapsed time between turn take up had to be 3 s or less, and a speaker could have no more than 3 s between vocalizations within a given turn. Given African

Grey parrots‘ predisposition to turn-taking, and the learned component associated with

turn-taking in both human and nonhuman species, we predicted Cosmo‘s vocal interactions with Betty Jean would mirror human turn-taking rules learned from BJ. This would include turn take-up times ~0.8-1.5 s and avoidance of interruption or simultaneous vocalizing.

In Chapters 4 and 5, I describe an alternate explanation for Cosmo‘s vocalization which states that Cosmo‘s vocalizations may serve more of a social function than a semantic or conversational function. To address this possibility, the current study expanded upon my previous work which showed that Cosmo used contextually- and thematically-linked vocalizations (Colbert-White et al., 2011). To do this, we placed

Cosmo‘s vocalizations back within the context of conversation with Betty Jean to establish the extent to which Cosmo maintained thematically cohesive conversations (i.e., dialogues). A dialogue was operationally defined as at least two turn-taking bouts of thematically-linked vocalizations, where theme adherence was subjectively assessed. If

Cosmo‘s vocalizations serve a primarily social, attention-getting function, there should be no evidence of dialogues in Cosmo‘s conversations. If dialogues are found, it would imply that Cosmo monitors Betty Jean‘s previous vocalizations and responds according to their thematic content, thereby supporting more advanced functions of Cosmo‘s repertoire.

Method

Subject and Housing

Cosmo was 6 years old at the beginning of the project in 2007. Betty Jean

(hereafter BJ) purchased Cosmo from a pet store in 2002 when she was 5 months old.

Although Cosmo had some experience hearing other human speakers, BJ was her

consistent companion. BJ established a simplified grammar with limited vocabulary by labeling new objects for Cosmo and correcting Cosmo‘s misuse and mispronunciation of words. Social interaction with BJ was the primary means by which Cosmo acquired melodies, English speech, and some nonword sounds like kiss noises. Additional nonword sounds such as microwave beeps were acquired naturally. Two female dogs also resided with BJ and Cosmo at the time of data collection.

All video-taping was conducted at BJ‘s home with Cosmo in her primary cage

(55.9 x 61.0 x 83.8 cm, with perch extending 40.6 cm from top) which was located in a sun room facing BJ‘s reading chair 2 m away. Food and water were provided ad libitum throughout all testing.

Recording and Transcriptions

For three weeks prior to testing, Cosmo was habituated to a Sony DCR-TRV39 mini-DV video camera on a tripod 1.5 m from the cage. Cosmo‘s cage was the only object in the camera frame. The camera‘s built-in microphone (32 kHz, 16-bit audio) recorded all audio. After the habituation phase, BJ recorded five, hour-long events for each of two social conditions during times when BJ thought Cosmo would be particularly talkative. In the IN condition, BJ sat in her reading chair and interacted with Cosmo as normal in the room with her. In the OUT condition, BJ remained in an adjacent room, but interacted with Cosmo as normal. An illustration of the layout of the space is shown in Figure 2.1.

decrease experimenter bias, contextual information in the videos was not used to construct the transcriptions. Nonword sounds were transcribed with two- or three-letter codes (refer to Table 2.1 for nonword sound coding scheme).

The IN (6 sessions) and OUT (6 sessions) transcriptions used for analysis contained 220 and 240 min, respectively, of usable footage. An independent observer transcribed approximately 5% of the tapes. ―Matching‖ required both timestamps and vocalizations to be identical between the transcribers. The Cohen‘s kappa coefficients of reliability for the seven reliability samples ranged from κ = .38-.79. When the .38 outlier was removed, the median kappa coefficient rose from .67 to .72.

Procedure

The time-stamped transcriptions were organized into one spreadsheet. During instances of vocal interaction between Cosmo and BJ, vocalization onset and offset times were recorded for both speakers. Originally, conversations were arbitrarily defined as at least two back-and-forths with fewer than 5 s turn take-up time between speakers. This time was reduced to 3 s after frequency distributions of turn take-up times during 6 randomly selected sessions (4 OUT, 2 IN) for both speakers demonstrated that there was a sharp drop off in turn take-up frequencies beyond 3 s (see Figures 2.2 and 2.3). This justified the use of 3 s in all subsequent conversation analyses.

Once conversations were identified, they were evaluated for thematic content to determine whether they were dialogues. This involved individually evaluating the

English meaning of the vocalizations in the conversations. Only those conversations which met the additional criterion of at least two back-and-forths of thematically-linked content were classified as dialogues.

Data Analysis

Assessments of conversation content were made via manual review of the data.

This included descriptions of instances of interruptions and both speakers‘ responses to interruptions, as well as the types of and themes of vocalizations that were used during turn-taking bouts.

Results

Turn-Taking

In the six randomly selected sessions, both speakers took turns vocalizing a similar number of times (330 and 332 for Cosmo and BJ, respectively). Cosmo took up the turn 0-56 s after BJ‘s vocalization offsets. BJ took up the turn 0-16 s following

Cosmo‘s offsets. BJ‘s median turn take-up time (1 s) was within Garvey and Berninger‘s

(1981) reported normal human range of ~0.8-1.5 s. Cosmo‘s median turn take-up time of

2 s was also very close to the human average.

Both frequency plots for turn take-up time showed non-normal distributions (see

Figures 2.2 and 2.3). Cosmo‘s turn take-up time distribution was strongly positively skewed (skewness of 5.07, SE = 0.29), where skewness > 1 is considered ‗highly skewed‘

(Bulmer, 1979). BJ‘s distribution was also strongly positively skewed (skewness of 4.67,

SE = 0.09). The distribution plot for BJ had much less variability, as demonstrated by most data points falling within the 0-3 s range.

Within the larger corpus of 12 taped sessions, turn-taking behavior was assumed to be similar relative to the 6 sessions used to conduct the above turn-taking analyses.

Despite similar turn take-up times, Cosmo and BJ behaved differently following silences equal to or greater than 20 s. Cosmo was numerically more likely to reinitiate contact

following silences (35 times). For 25 of Cosmo‘s 35 reinitiations following silences, the last utterance prior to the silence was made by Cosmo; for the remaining 10, BJ was the last speaker. In contrast, BJ only reinitiated contact following silences 9 times. Five reinitiations were following BJ as the last speaker before the silence; 4 were following

Cosmo as the last speaker. Due to small ns, these data could not be explored statistically.

Conversations

According to the definition of at least two turns with ≤ 3 s between speakers and ≤

3 s between vocalizations within a speaker‘s turn, there were 124 conversations distributed across 11 of the 12 taped sessions. Eighty-five conversations were distributed across 6 IN sessions, and 39 conversations were distributed across 5 OUT sessions. The number of conversations when BJ was in the room with Cosmo compared to when they were visually separated was found to be different from what would be expected by chance, X2(1) = 91.32, p < .001. Table 2.2 provides descriptive data for each conversation.

Each session contained between 2 and 22 conversations (Mean = 11.27, SD =

7.30). On average, there were numerically more conversations per session in the IN condition than in the OUT condition, though a Student t-test assuming unequal variances revealed this difference was not significant, t(9) = -1.58, p = .15. Lengths of conversations ranged from 6-81 s (Mode = 15 s, Median = 20 s, Mean = 23.48 s, SD =

15.75). IN session conversations ranged from 6-78 s, with a mean conversation length of

23.05 s (SD = 15.76); OUT conversations ranged from 7-81 s, with a mean of 23.67 s (SD

= 15.83), indicating conversation length was unaffected by the location of the speakers.

A frequency distribution of conversation lengths is provided in Figure 2.4.

The number of turn-taking bouts (i.e., coupled BJ-Cosmo or Cosmo-BJ utterances) ranged from 2-9 (Median = 3, Mode = 2, Mean = 3.56 bouts, SD = 1.94). The mean number of bouts taken when BJ was in the room with Cosmo (3.66 turns, SD =

1.92) was very similar to the mean number of bouts when the two speakers were separated (3.31 turns, SD = 1.52). A frequency distribution of turn-taking bouts is provided in Figure 2.5.

There was a strong bias for Cosmo to initiate conversations, both while she was in the room with BJ and when they were in separate rooms. In the IN condition, Cosmo initiated 73 of the 85 conversations; in the OUT condition, Cosmo initiated 38 of the 39 conversations. The few conversations which BJ did initiate fell into three main categories: Contact (e.g., DUW [Duet whistle]), Affection (e.g., ―I love you‖), and

Labeling (e.g., ―That‘s squirrel‖). A summary of BJ‘s initiations is provided in Table 2.3.

A chi-square test confirmed that frequencies of initiation were not equal between the speakers or the two conditions, X2(1) = 14385, p < .001.

In the 124 conversations, there were 45 instances of interruption. Cosmo interrupted BJ 28 times; BJ interrupted Cosmo 17 times. The distribution of interruptions between the two speakers was unequal from what would be expected by chance, X2(1) =

16.34, p < .001. There was no observed pattern in the content of which utterances were interrupted, by whom, or which utterances were interjected during interruptions. Most interruptions, however, subjectively appeared to be when one speaker had not finished her turn and the other took over, as with both speakers overlapping each other‘s DUW

[Duet whistle]. BJ also interrupted Cosmo‘s nonword sounds (e.g., NWM

[Miscellaneous nonword sound], DUW, RI [Telephone ring], WF [BJ‘s dog bark sound])

with laughter, DUW, or speech such as ―What‘s that?‖ In only 3 out of 28 instances did

Cosmo interrupt BJ following BJ refusing one of Cosmo‘s requests. During two of these times Cosmo said ―No‖ back to BJ and in one instance she interrupted with ―Be a good.‖

For a given speaker, a turn could be either a single stream of vocalizations or a series of multiple streams. Very few conversations were comprised completely of single- stream back-and-forth bouts (see Table 2.4 for examples of single-stream vs. multiple- streams). Of the 85 conversations while BJ and Cosmo were in the same room together, only 12 contained single-stream back-and-forth duetting. Nine had two single-stream back-and-forth bouts, two had three single-stream back-and-forths, and one had five. Of the 39 conversations in the OUT of the room condition, four had two single-stream back- and-forth bouts, two had three single-stream back-and-forths, and one had eight (see

Table 2.5). More common were duets in which speakers uttered multiple streams within the ≤ 3 s per turn criterion.

When both speakers were in the room, Cosmo vocalized up to 8 vocal streams within one turn (see Table 2.4), while BJ uttered no more than 5 vocal streams within one turn. When both speakers were out of the room, Cosmo uttered as many as 9 vocal streams within 3 s of each other before BJ responded. In comparison, BJ never uttered more than 2 vocal streams within one turn, and this only occurred 5 times. Single-stream turn-taking bouts were most common during short conversations (i.e., < 4 bouts of turn- taking). The content in these single-stream conversations varied, but some centered around the theme of contact as with one brief conversation initiated by Cosmo, ―I‘m here‖ ―There you are Cosmo‖ ―Here you are‖ ―DUW.‖

Dialogues

Overall, the thematic content of the conversations varied considerably both between the two social conditions, as well as within individual conversations themselves.

While BJ‘s responses to Cosmo‘s vocalizations were thematically appropriate, Cosmo did not always maintain themes. For example, one conversation appeared thematically- linked with both speakers taking turns duet-whistling back and forth until Cosmo said,

―Betty Jean have go in a car,‖ when there had been no prior utterances from BJ or Cosmo about BJ going in a car. Further, neither speaker pursued this theme after the utterance was made, making it even more thematically inappropriate.

When conversations were broken down further into dialogues (defined as at least two turn-taking bouts of thematically-linked vocalizations), thematically appropriate patterns did emerge. Out of 124 conversations, only 32 (26%) met the criteria of having thematically-linked content throughout the conversation. For those conversations which were dialogues, thematic content is provided in Table 2.2. The 11 common themes and their frequency of use within the 32 dialogues appear in Table 2.6. Dialogues pertained largely to vocal contact through DUW (10 times) or contact requests to whistle or talk

(10 times). Affection (specifically kisses) was also a common theme of dialogues (8 times). Surprisingly, requesting to get out of the cage—which is addressed further in

Chapter 4—was not a frequently appearing theme, and appeared in only 4 of the 32 dialogues.

Discussion

The current study extensively described, for the first time, spontaneous vocal interactions between a human and nonhuman speaker. In particular, we looked for

evidence of the human conversation strategies of accepted turn take-up times, cohesiveness of conversation content, and interruption. This was done to assess whether it is possible for a nonhuman to learn and incorporate human conversation strategies into its vocal interactions with a human social partner.

Similar to human infants as young as 3 months (Bloom et al., 1987), Cosmo adhered to human turn-taking rules. This is in accordance with observations of turn- taking in wild African Greys (unpublished data, Pepperberg, 1999). Additionally,

Cosmo‘s median turn take-up time of 2 s fell very close to average human turn take-up time (~0.8-1.5 s, Garvey & Berninger, 1981). This suggests either that parrots and humans independently converged upon similar turn take-up times, or, more likely, Cosmo has learned from BJ what constitutes the end of a turn, a pause within a partner‘s turn, or an unresponsive partner. While future research would be necessary to test Cosmo‘s response to these three types of silences, positive findings would be particularly exciting because they would show that a nonhuman speaker can flexibly learn to use at least one human conversational strategy.

While both speakers interrupted one another, we were surprised to find that

Cosmo did not appear to interrupt following refusals by BJ. Young children often respond to refusals with frustration (Schiffrin, 1985), and interruption during conversation is a behavioral indicator of frustration (Gallagher & Craig, 1981).

However, interruptions by Cosmo seemed largely coincidental, such as when BJ had not completed a turn and Cosmo began her turn. For Cosmo, it could be that being refused is not associated with frustration—as we would know it in humans—because requests

(refused or met) could serve a pragmatic purpose of simply prolonging interaction with

BJ. Requests made to BJ consistently solicit a response; whereas other vocalizations such as answering machine beeps do not. Alternatively, Cosmo may not express frustration through vocal communication the way that humans do.

The possibility that many of Cosmo‘s vocalizations serve a pragmatic rather than literal function is one which will be explored further in subsequent chapters. Many species which engage in antiphonal duetting do so for pair-bond maintenance functions

(Douglas & Mennill, 2010). That is to say, specific vocalizations are combined and uttered to serve a holistic function, rather than being used discretely. Cosmo was much more likely to initiate and reinitiate contact with BJ, regardless of whether she was in the room with BJ or visually separated. Since many bonded parrots prefer to remain in physical contact with their mates (Seibert, 2006), given Cosmo‘s confinement in a cage, maintaining vocal contact may be a surrogate form of social interaction. This could explain her frequent initiation and reinitiation of contact with BJ, as well as her pattern of taking multiple turns within turns when BJ did not respond immediately.

Interestingly, our findings that Cosmo initiated conversation with BJ using her artificial communication system oppose the findings of at least one famous nonhuman primate language study. Terrace‘s (1986) highly debated conclusions about his signing chimpanzee Nim Chimpsky were that Nim rarely initiated conversation with his handlers, frequently only repeated what his handlers signed to him first, and did more requesting than conversing. Conversations between Cosmo and BJ were much more human-like in many ways. Cosmo frequently initiated conversation, responded to BJ‘s vocalizations variably depending upon the social or speech context, and held thematically-linked

conversations which did not always pertain to requests (e.g., BJ going to work or having company).

Only 26% of conversations between Cosmo and BJ were considered thematically cohesive enough to be called dialogues. While this number does not mirror normal adult conversations, or even the conversations of normally-developing children as young as 4 years (Adams & Bishop, 1989), the results are still promising as they may address questions regarding both speech-using and wild African Greys‘ vocal communication systems. Pepperberg conducted numerous studies with her African Grey Alex after training him using her intensive Model/Rival (M/R) paradigm (see Pepperberg, 1981,

1988; Todt, 1975). Pepperberg argued that M/R training was responsible for Alex‘s referential label-learning abilities. According to BJ, Cosmo never had formal speech training (personal communication, 2009). Instead, BJ used speech and nonword sounds to interact with Cosmo spontaneously, though she did label objects and correct Cosmo‘s misuse and mispronunciation of words. Though there are no published transcriptions of spontaneous conversations between Alex and Pepperberg, we believe differences in the way Alex and Cosmo developed their repertoires (training of discrete words vs. no training) might explain why Cosmo‘s repertoire appears to be more socially-driven in its use. That is to say, Cosmo‘s repertoire may be based heavily upon those utterances which initiate or prolong conversation, whereas Alex‘s repertoire use is a reflection of his discrete vocalization training. Comparisons of transcriptions may empirically support this hypothesis.

In the wild, many parrots engage in antiphonal duetting, and some have specific vocalizations for maintaining social contact during times of visual separation (e.g.,

Nottebohm, 1972; Serpell, 1981; see Colbert-White et al., 2011 for evidence of contact calls in a home-reared parrot). Given the strong ecological relevance of duetting to parrots, it is no surprise that Cosmo exhibits this species-typical behavior with BJ. What is interesting, however, is the variation and lack of cohesion in many of the conversations. Some parrots have specific calls which are unique to a particular pair and serve the function of contact and bonding (e.g., yellow-naped Amazon, Amazona auropalliata, Wright, 1996). While Cosmo and BJ have such ―signature‖ pair vocalizations where either partner can produce all of the parts (e.g., the DUW whistle of

Bridge Over the River Kwai), both Cosmo and BJ also went back-and-forth with many novel phrases. Even more, within these novel duets, the next speaker‘s utterance was sometimes related thematically to the previous utterance of the partner. Cosmo‘s novel duetting behavior indicates that wild paired parrots may also produce utterances generatively and flexibly in their duets. Such a finding would be of great value to those studying animal communication systems in the wild and in captivity.

Acknowledgements

Hannah Hall and Marisol Macias both aided with literature searching. In addition, I thank Hannah Hall and Allison Kaufman for assisting with reliability coding and associated calculations, respectively, and Dominic Byrd for his illustration of the testing area in Figure 2.1.

References

Adams, C., & Bishop, D. V. M. (1989). Conversational characteristics of children with

semantic-pragmatic disorder. I: Exchange structure, turntaking, repairs and

cohesion. British Journal of Disorders of Communication, 24, 211-239.

Bloom, K., Russell, A., & Wassenberg, K. (1987). Turn taking affects the quality of

infant vocalizations. Journal of Child Language, 14, 211-227.

doi:10.1017/S0305000900012897

Bottoni, L., Massa, R., & Boero, D. L. (2003). The grey parrot (Psittacus erithacus) as

musician: An experiment with the Temperate Scale. Ethology Ecology &

Evolution, 15, 133-141. doi:10.1080/08927014.2003.9522678

Bulmer, M. G. (1979). Principles of statistics. Chelmsford, MA: Courier Dover.

Colbert-White, E. N., Covington, M. A., & Fragaszy, D. M. (2011). Social context

influences the vocalizations of a home-raised African Grey parrot (Psittacus

erithacus erithacus). Journal of Comparative Psychology, 125, 175-184.

doi:10.1037/a0022097

Craig, H. K., & Gallagher, T. M. (1983). Adult-child discourse: The conversational

relevance of pauses. Journal of Pragmatics 7(4), 347-360.

Douglas, S. B., & Mennill, D. J. (2010). A review of acoustic playback techniques for

studying avian vocal duets. Journal of Field Ornithology, 81, 115-129.

doi:0.1111/j.1557-9263.2010.00268.x

Gallagher, T. M., & Craig, H. K. (1982). An investigation of overlap in children‘s

speech. Journal of Psycholinguistic Research, 11, 63-75.

doi:10.1007/BF01067502

Garvey, C., & Berninger, G. (1981). Timing and turn taking in children‘s conversations.

Discourse Processes, 4, 27-57. doi:10.1080/01638538109544505

Haimoff, E. H. (1986). Convergence in the duetting of monogamous Old World primates.

Journal of Human Evolution, 15, 51-59.

Lemasson, A., Glas, L., Barbu, SD., Lacroix, A., Guilloux, M., Remeuf, K., & Koda, H.

(2011). Youngsters do not pay attention to conversational rules: Is this so for

nonhuman primates? Scientific Reports, 1, 1-4. doi:10.1038/srep00022

Méndez-Cárdenas, M. G., & Zimmermann, E. (2009). Duetting—A mechanism to

strengthen pair bonds in a dispersed pair-living primate (Lepilemur edwardsi)?

American Journal of Physical Anthropology, 139, 523-532.

doi:10.1002/ajpa.21017

Müller, A. E., & Anzenberger, G. (2002). Duetting in the Titi Monkey (Callicebus

cupreus): Structure, pair specificity and development of duets. Folia

Primatologica, 73, 104-115. doi:10.1159/000064788

Nottebohm, F. (1972). The origins of vocal learning. The American Naturalist, 106(947),

116-140.

Paltridge, B. (2006). Discourse analysis. London: Continuum.

Pepperberg, I. M. (1981). Functional vocalizations by an African Grey parrot (Psittacus

erithacus). Zeitschrift fürTierpsychologie, 55, 139-160. doi: 10.1111/j.1439-

0310.1981.tb01265.x

Pepperberg, I. M. (1988). An interactive modeling technique for acquisition of

communication skills separation of labeling and requesting in a psittacine subject.

Applied Psycholinguist, 9, 59-76.

doi:http://dx.doi.org/10.1017/S014271640000045X

Pepperberg, I. M. (1999). The Alex studies: Cognitive and communicative abilities of

Grey parrots. Cambridge, MA: Harvard University Press.

Rutter, D. R., & Durkin, K. (1987). Turn-taking in mother-infant interaction: An

examination of vocalizations and gaze. Developmental Psychology, 23, 54-61.

doi:10.1037/0012-1649.23.1.54

Schegloff, E. A. (1968). Sequencing in conversational openings. American

Anthropologist, 70, 1075-1095.

Schiffrin, D. (1985). Everyday argument: The organization of diversity in talk. In: D.

Tuen, & A. Van (Eds.), Handbook of discourse analysis: Discourse and dialogue,

(pp. 35-46). London: Academic Press.

Serpell, J. (1981). Duets, greetings and triumph ceremonies: Analogous displays in the

parrot genus Trichoglossus. Zeitschrift Tierpsychologie, 55, 268-283.

doi:10.1111/j.1439-0310.1981.tb01272.x

Sonnenschein, E., & Reyer, H. U. (1983). Mate-guarding and other functions of

antiphonal duets in the slate-coloured boubou (laniarius funebris). Zeitschrift für

Tierpsychologie, 63, 112-140. doi:10.1111/j.1439-0310.1983.tb00083.x

Snowdon, C. T., & Cleveland, J. (2005). ―Conversations‖ among pygmy marmosets.

American Journal of Primatology, 7, 15-20. doi:10.1002/ajp.1350070104

Thalmann, U., Geissmann, T., Simona, A., & Mutschler, T. (1993). The indris of

Anjanaharibe-Sud, northeastern Madagascar. International Journal of

Primatology, 14, 357-381. doi:10.1007/BF02192772

Todt , D. (1975). Social learning of vocal patterns and modes of their applications in

Grey parrots (Psittacus erithacus). Zeitschrift fürTierpsychologie, 39, 178-188.

doi:10.1111/j.1439-0310.1975.tb00907.x

Todt., D., & Fiebelkorn, A. (1980). Display, timing and function of wing movements

accompanying antiphonal duets of Cichladusa guttata. Behaviour, 72(1/2), 82-

106.

Table 2.1.

Nonword Sound Coding Scheme

CODE DESCRIPTION CODE DESCRIPTION

AM Telephone answering machine beep LS Laser sound

CR Crow caw LSW ―Laser sound-ID-whistle‖ combination

DS Any dog bark, gruff, or howl MWH Miscellaneous one-note whistle

DSS Any dog bark, gruff, or howl sequence NWM Miscellaneous non-whistle

DO Door opening creak NWMS Miscellaneous non-whistle sequence

DOS Door opening creak sequence OOO ―Ooh‖ and other long-o sounds

DUW Duet whistle OU ―Oww‖ as in pain

DUWS Duet whistle sequence OW Owl hoot

DW Dog whine/whimper PH Telephone dialing beep

DWS Dog whine/whimper sequence PHS More than five telephone dialing beeps

FR Frog croak RI Telephone ring

HA Hawk vocalization RIS Telephone ring sequence

HAS Hawk vocalization sequence WBI Wild miscellaneous songbird

ID Indistinguishable WBIS Wild miscellaneous songbird sequence

―Woo-woo-woo‖ or ―woo‖ (B. J.‘s KS Kiss sound WF vocalization of a dog barking)

KSS Kiss sound sequence WW Wolf whistle

LA Laugh WWS Wolf whistle sequence

Note. ―Sequence‖ = Vocalization repeatedly uttered for more than 4 seconds. ―Duet whistle‖ = Melodic whistling, not natural bird vocalization.

Table 2.2.

Length, Turn, Interruption, and Theme Data for Each Conversation

Video Onset Time Length of Number of Interruption Theme of Dialogue (If

Conversation (s) Coupled Turns (By Whom?) Applicable)

OUT110907 41:06:00 36 4 Label-learning WF

46:16:00 17 2 BJ impressed by Cosmo‘s WW

OUT052908 2:50 23 3

4:26 14 3

16:56 11 3 Cosmo Telephone games

22:33 23 4 BJ

26:37:00 30 4 B J

28:43:00 11 2

29:13:00 44 4 DUW; Contact; ―I love you‖ ; KS

31:43:00 30 3 BJ

33:21:00 12 2

33:37:00 7 2

OUT012108 6:08 29 4 Location; ―I‘m here‖; Contact

7:09 25 3

13:32 21 4 Cosmo Telephone games

20:03 21 4

22:15 16 4 DUW; ―I‘m here‖; Location; Contact

26:28:00 8 2 Location; Contact; ―I‘m here‖

29:56:00 41 4

31:43:00 19 3 DUW; Contact

32:26:00 20 4

34:10:00 22 5 BJ

34:55:00 62 7 Cosmo

36:11:00 44 6

39:45:00 17 3

40:30:00 10 2

41:21:00 12 2 Cosmo Contact; DUW

OUT011308 20:15 14 2

26:08:00 34 4

OUT111807 19:27 7 2

20:01 29 3 Cosmo

20:52 40 4 Cosmo DUW; Location; Contact

22:46 81 9

50:49:00 10 2 Cosmo

53:51:00 8 2

56:42:00 11 2 DUW

57:44:00 15 2

58:25:00 16 2 Cosmo

59:03:00 9 2

IN110907 6:12 15 3 ―Go up‖ (i.e., get out of cage)

6:48 15 3 Cosmo ―Be a good bird‖; ―Go up‖

7:14 36 5 Cosmo (1x),

BJ (2x)

10:58 35 8 Cosmo

11:41 65 9 BJ

16:16 24 4

16:53 12 2

17:59 15 3

18:44 16 3

19:27 13 3

19:51 37 6

25:14:00 14 3

26:14:00 39 6 Cosmo

26:28:00 25 4 Cosmo I love you; DUW

27:13:00 9 2 Label-learning ―television‖

27:31:00 65 9 Label-learning ―television‖; Telephone

28:42:00 10 2

IN102107 49:14:00 29 4

50:52:00 44 5 kiss, cuddle, love you

58:36:00 12 2

IN120907 21:02 12 2

22:01 16 3 Greetings; KS

22:23 9 2 ―I love you‖; Cuddle

23:43 16 2

24:48:00 8 2

25:17:00 12 2

25:55:00 26 3

27:57:00 10 2 BJ‘s dogs (WF, ―Kerri‖)

28:40:00 31 5

29:29:00 21 2 Contact; DUW

31:10:00 15 2

31:30:00 20 2

34:20:00 8 2 BJ

35:10:00 26 4 Cosmo ―Cosmo be a good bird‖; Go up

36:14:00 11 2

36:32:00 28 4

46:53:00 15 3

47:43:00 16 3 Cosmo

48:08:00 17 3 BJ (2x)

48:30:00 33 3

49:42:00 38 5

IN110407 4:05 23 3 BJ (2x)

4:34 73 7 Cosmo (1x),

BJ (1x)

5:47 21 4 Cosmo

6:36 44 4 Cosmo (2x)

7:25 59 6

8:32 36 6 DUW; Affection; Contact

13:25 16 3

13:55 78 9

15:17 7 2 BJ

16:53 63 9 Cosmo

19:21 9 2

20:19 15 2 Greeting; telephone; ―Hi Tom‖

21:53 10 2

26:19:00 7 2 Cosmo Greeting; DUW

27:49:00 28 4 Greeting; Contact

29:12:00 14 2

30:10:00 53 9 Telephone games

31:08:00 11 2 Cosmo

31:21:00 14 3

31:56:00 8 2

32:43:00 13 3 Kiss; KS

44:37:00 25 2 BJ

IN011308 2:47 17 4

3:12 24 5 Cosmo

4:17 8 2

4:34 7 2 Cosmo

4:45 22 5 Cosmo

6:26 6 2

12:09 24 5

40:06:00 33 3

40:43:00 20 3

41:33:00 27 4 BJ

42:08:00 28 4

43:57:00 15 4 Cosmo

44:38:00 27 3

45:13:00 29 4 Cosmo

58:22:00 40 7 Cosmo

IN011908 28:07:00 28 5 Cosmo, BJ ―Wanna come here‖; ―Go up‖

29:35:00 17 3 Kiss; Come here

32:58:00 20 2

33:33:00 20 3

33:58:00 25 4 Fur; Feathers; Clothes

35:19:00 6 2

37:40:00 24 3

Table 2.3.

Conversation Initiations Made by Betty Jean

DUW [Duet whistle] (2x)

―I love you (3x)‖

KS [Kiss sound]

―What?‖ following > 3 s pause after Cosmo uttered ―You have reached‖

―Telephone for bird‖ following > 3 s pause acknowledging Cosmo‘s telephone sound

―Cosmo has feet‖

―Cosmo stay in cage okay?‖ following > 3 s pause after telling Cosmo she wanted to read the newspaper rather than let Cosmo out of her cage

―That‘s squirrel‖ labeling for Cosmo after showing her a squirrel outside

―Television‖ following > 3 s pause after working with Cosmo to say ―Television‖

LA [Laugh sound] following a joke made by Cosmo > 3 s prior

Note. Question marks were added post-hoc to illustrate instances of upward tonal inflection. Original transcriptions contained no punctuation. Turns which were not taken within 3 s were excluded from consideration as being part of a conversation.

Table 2.4.

Single and Multiple Vocal Streams Within Turns

Excerpt A

Cosmo: Hello

BJ: Hi

Cosmo: How are you?

BJ: Fine thank you

Cosmo: Cos

BJ: How are you?

Cosmo: Wanna be a good bird?

BJ: Yes

Excerpt B

Cosmo: Oh

Cosmo wanna peanut?

Okay

Here I are

MWH

I‘m here

Here you are

BJ: I‘m here

Note. Question marks were added post-hoc to illustrate instances of upward tonal inflection. Original transcriptions contained no punctuation.

Table 2.5.

Longest Single-Turn Turn-Taking Event

Cosmo: Look

BJ: What?

Cosmo: Cosmo wanna talk

BJ: Okay let‘s talk

Cosmo: Betty Jean wanna whistle DUW

BJ: DUW

Cosmo: DUW

BJ: DUW

Cosmo: DUW

BJ: DUW

Cosmo: Cosmo wanna whistle DUW

BJ: DUW

Cosmo: Cosmo and Betty Jean wanna whistle

BJ: Okay

Cosmo: DUW

BJ: DUW

Cosmo: DUW

Table 2.6.

Common Themes in Dialogues and Their Respective Frequencies

THEME FREQUENCY

DUW [Duet whistle] 10

Contact (Requests to whistle or talk) 10

Affection (e.g., ―I love you‖, KS [Kiss], requests for kisses) 8

Telephone (e.g., Phone dialing, ―Telephone for bird‖) 4

Greetings (e.g., ―Hello,‖ ―Hi how are you‖) 4

Go up (Requests to get out of cage) 4

Location (e.g., ―Where are you‖ ―I‘m here‖) 3

Promises (―Don‘t bite‖ or ―Be a good bird‖) 2

Label-learning (e.g., ―What‘s WF‖ ―Television‖) 2

Dogs (e.g., ―Hello Kerri,‖ ―Doggies have fur‖) 2

Come here (Requests for BJ to come to Cosmo‘s cage) 2

Figure 2.1. Illustration of the testing space, including relevant dimensions and furniture.

Figure 2.2. Frequency plot for Cosmo‘s turn take-up time. The last category, ‘20-56‘ contained six scores (20, 22, 29, 30, 34, and 56).

Figure 2.3. Frequency plot for BJ‘s turn take-up times.

Figure 2.4. Frequency plot for conversation length. Horizontal axis represents bins of 5 s (i.e., 0-5 s, 6-10 s, etc).

Figure 2.5. Frequency plot for coupled turns, or ―turn-taking bouts,‖ where one speaker vocalizes and the other responds in 3 or fewer seconds. The definition also included a limitation to 3 s for the pause length between utterances by one speaker within a turn.

CHAPTER 3

VOLUME MODULATION AS EVIDENCE FOR AUDITORY PERCEPUTAL

PERSPECTIVE TAKING BY AN AFRICAN GREY PARROT1

______

1 Colbert-White, E. N., to be submitted

Abstract

In humans, auditory perceptual perspective taking (i.e., the ability to take the perspective of what another individual can hear) develops around age 5. While many species modify their vocalizations (e.g., changing rate, pitch) in response to environmental noise, only a few studies have supported the possibility of perceptual perspective taking in nonhumans.

The current study explored one communication boundary, distance, and examined the vocal behavior of one 6-year-old African Grey parrot. The parrot‘s repertoire units were dichotomously categorized as being either social (e.g., ―Wanna kiss‖) or nonsocial (e.g., beeping sounds). The parrot‘s owner was asked to sit in an adjacent room out of visual contact, and conversations between the two speakers were recorded. As predicted, the parrot vocalized significantly more loudly when her owner was out of the room compared to baseline conversations between her and her owner in the same room. Further, social vocalizations were uttered significantly more loudly than nonsocial utterances. The results indicate that at least one African Grey parrot may modify volume in response to the auditory barrier of distance, a possible indicator of perceptual perspective taking.

Keywords: African Grey parrot, Psittacus erithacus, perceptual perspective taking, language, communication

Introduction

Throughout the animal kingdom, spontaneous modification of qualities of vocalizations (e.g., pitch, duration, volume, rate) ensures accurate information transfer from sender to receiver. For humans and nonhumans alike, qualities of vocalizations can provide social information beyond the vocalizations being uttered. In humans, this can include details about a speaker‘s emotional state (i.e., positive or negative) or level of confidence, and can induce emotions like arousal in listeners (Aeschlimann, Knebel,

Murray, & Clarke, 2008). Some vocalization modifications (e.g., tone of voice) can even dictate the meaning or intended interpretation of a vocalization. This is the case in tonal languages such as Chinese and in linguistic devices such as irony, respectively (Kreuz &

Roberts, 1995).

Individual differences in the qualities of nonhumans‘ vocalizations are usually involuntary, honest signals of fitness. For example, in the frigatebird (Fregata minor), vocalization frequency (measured in kHz) correlates strongly with body size and mating potential (Juola & Searcy, 2011). However, many species voluntarily modify qualities of their vocalizations such as rate and volume in the presence of social partners (e.g., alarm calls in bobwhite quail, Colinus virginianus, Evans & Marler, 1991; food calls in chickens, Gallus gallus, Evans & Marler, 1994). It is presumed that intentional modifications by some nonhumans serve a similar purpose to those made by humans— that is, to provide additional social information about the sender‘s state. In the case of some cognitively advanced species, quality modifications are even argued to indicate details about mood or emotional state (e.g., dolphins, g. Tursiops, Caldwell & Caldwell,

1977).

In humans, the quality of volume is modified under a variety of circumstances.

Decreasing or increasing vocalization volume during conversation can signal the end of a turn or the opportunity to take over the turn during a conversation (Paltridge, 2006).

From a young age, children learn that raising the voice is a predictable indicator of anger during conflict (Esling, 1994); while reducing volume is an indication that the information being transferred is private. Though research has shown that other species spontaneously modify volume and respond to changes in conspecifics‘ vocalization volume, the role volume modification plays in bird vocalizations has been studied minimally in comparison to qualities such as song tempo and frequency modulation

(Pasteau, Ung, Kreutzer, & Aubin, 2012). Further, even less is known about volume (or any other quality) modification in members of the parrot family.

The modification of volume is particularly relevant when humans must communicate across auditory or visual barriers. That is to say, humans take into account factors such as distance and whether the listener‘s hearing is compromised (e.g., listener is wearing headphones), then modify their vocalizations accordingly. This ability arises as children develop theory of mind, a skill-set of cognitive abilities which, when combined, allow individuals to attribute mental states to others (Premack & Woodruff,

1978). Perspective taking can be divided into two categories: cognitive and perceptual

(some include a third, affective perspective taking, e.g., Kurdek & Rodgon, 1975). By around age 6, children are able to engage in cognitive perspective taking; that is, recognizing that their own intentions, desires, and beliefs can be different from others‘

(Axia, 1996; Clark & Delia, 1976). For example, a child who is denied the opportunity to go to the zoo might retort by saying, ―But you like gorillas, and we would get to see

gorillas if we went to the zoo.‖ In perceptual perspective taking, children take on another‘s perceptions, as in Piaget‘s (1967) classic three mountain problem. In the task, children were seated opposite a doll with a mountain landscape between them and asked to report what objects the doll saw on the mountain. Children under age 5 consistently reported objects visible on their side of the mountain, while older children walked around to the doll‘s side before responding. While this and other perceptual perspective taking problems (e.g., Kurdek & Rodgon, 1975; Borke, 1975; Michelon & Zacks, 2006) involve taking the perspective of another‘s visual field, there is very little reporting children assuming a social partner‘s olfactory, gustatory, tactile, or, most relevant to this study, auditory perspective.

Among nonhumans, there are two camps of thought with respect to modification of vocalizations. Some interpret the modification as a learned behavioral process related to mating and territory defense. For example, male leptodactylid frogs, Eupsophus calcaratus, responded to simulated background noise (i.e., artificial wind, sea surf, or conspecific calls) by increasing their call rate and duration (Penna, Pottstock, &

Velasquez, 2005), which the authors concluded was ensured accurate information transfer over the simulated noise. A similar interpretation has been made of urban birds modifying their vocalizations in response to ambient city noise. Slabbekoorn and Peet

(2003) recorded urban male great tits (Parus major) in their territories and correlated their song pitch with the territories‘ ambient noise level. They found that populations in louder environments sang at higher minimum frequencies (measured in kHz), and concluded that the adjustment was a learned mating strategy. In a follow up to

Slabbekoorn and Peet‘s work, Hu and Cardoso (2010) showed that only species with

intermediate species-typical minimum frequencies (~1.0-1.5 kHz) modified their vocalizations in response to ambient noise. Those with low minimum frequencies (i.e., below 1.0 kHz) did not. The authors hypothesized that for this second group, increasing minimum frequencies may not provide enough of an adjustment to overcome urban noise, and so those species might modify in other ways (e.g., more loudly, in between periods of urban noise). Rather than concluding that the animals engaged in perspective taking, the authors of the above three studies described the observed modification as evidence for selective behavioral plasticity (i.e., microevolution).

In contrast, the second camp of thought interprets certain modifications of behavior as positive evidence for perspective taking by nonhumans. A variety of caching and food retrieval tasks have shown that individuals of some species prefer to act when a conspecific predator is unaware (e.g., chimpanzees, Pan troglodytes, Melis, Call, &

Tomasello, 2006; rhesus macaques, Macaca mulatta, Santos, Nissen, & Ferrugia, 2006;

Western scrub-jays, Aphelocoma californica, Stulp, Emery, Verhulst, & Clayton, 2009; dogs, Canis familiaris, Kundey et al., 2010; ravens, Corvus corax, Bugnyar & Heinrich,

2005). This same perspective attribution has been demonstrated interspecifically with humans and horses (Equus caballus, Proops & McComb, 2010), humans and dolphins

(Tursiops truncatus, Xitco, Gory, & Kuczaj, 2001), humans and dogs (Call, Bräuer,

Kaminski, & Tomasello, 2003; Bräuer, Call, & Tomasello, 2004), and humans and

African Grey parrots (Psittacus erithacus, Péron, Rat-Fischer, Nagle, & Bovet, 2010;

Péron, Chardard, Nagle, & Bovet, 2011).

Parrots are highly social with complex relationship dynamics (Emery, Seed, von

Bayern, & Clayton 2007, Seibert, 2006) and a large brain to body size ratio (Emery et al.,

2007; Burish, Kueh, & Wang, 2004). They also provide extensive post-natal care (e.g.,

Bucher, 1983) and are predominantly monogamous (Seibert, 2006). These same traits can be found in humans, making parrots a strong candidate for demonstrating sociocognitive abilities related to partner monitoring. Further, as Péron et al. (2011) demonstrated, at least one species of parrot has demonstrated visual perceptual perspective taking. In the task, Péron and colleagues‘ African Grey parrots were provided with unsafe ―toys‖ (e.g., ink pens) then made to believe their handler was either present or not present. At least one of the subjects was more likely to go for the forbidden objects when the handler was believed to be not present. The authors concluded that the individual‘s behavior was indicative of visual perceptual perspective taking.

Given Péron and colleagues‘ (2011) perspective taking results, as well as Greys‘ well-documented cognitive abilities (e.g., Aïn, Giret, Grand, Kreutzer, & Bovet, 2009;

Giret et al., 2010; Péron et al., 2011; Péron, Rat-Fischer, Lalot, Nagle, & Bovet, 2011; see Pepperberg, 1999 for extensive review), the current study explored an alternative perceptual perspective-taking task, auditory perspective taking. One Grey was recorded during spontaneous conversations with her human caregiver at two distances. I predicted the parrot would vocalize significantly more loudly when she was visually separated from her owner (i.e., an auditory barrier) than when her owner was in the same room with her.

In addition, when the parrot and her owner were visually separated, vocalizations classified ad hoc as social (e.g., ―I love you‖) should be vocalized significantly more loudly than nonsocial (e.g., beeping sounds) vocalizations. Such findings might represent evidence that the bird was engaging in auditory perceptual perspective taking.

Method

Subject and Housing

Cosmo, a female Congo African Grey parrot (Psittacus erithacus), was the subject of the investigation. Cosmo was 6 years old at the beginning of the project in

2007. Cosmo‘s female owner Betty Jean (hereafter BJ) purchased her from a pet store in

2002 when she was 5 months old. Although Cosmo had some experience hearing other human speakers, BJ was her consistent companion. BJ established a simplified grammar with limited vocabulary by labeling new objects for Cosmo and correcting Cosmo‘s misuse and mispronunciation of words. Social interaction with BJ was the primary means by which Cosmo acquired melodies, English speech, and some nonword sounds like kiss noises. Additional nonword sounds such as microwave beeps were acquired naturally. Two female dogs also resided with BJ and Cosmo at the time of data collection.

All video-taping was conducted at BJ‘s home with Cosmo in her primary cage

(55.9 x 61.0 x 83.8 cm, with perch extending 40.6 cm from top) which was located in a sun room facing BJ‘s reading chair 2 m away. Food and water were provided ad libitum throughout all testing.

Recording and Transcriptions

In the AL (i.e., alone) condition, BJ started the video camera then left the house. In the IN condition, BJ sat in her reading chair and interacted with Cosmo as normal in the room with her. In the OUT condition, BJ remained in an adjacent room, but interacted with

Cosmo as normal. An illustration of the layout of the space is shown in Figure 3.1.

Videos were transcribed using the code ―ID‖ (i.e., indistinguishable) whenever vocalizations were not clear. Syllables and fragments were transcribed as they were heard (e.g., ―tele‖ and ―showe‖ as in ―telephone‖ and ―shower,‖ respectively). To decrease experimenter bias, contextual information in the videos was not used to construct the transcriptions. Nonword sounds were transcribed with two- or three-letter codes (refer to Table 3.1 for nonword sound coding scheme). While Colbert-White et al.

(2011) excluded units which were uttered only once, due to the small dataset, all units were included in the current study. Fragments (e.g., ―Mary is a doggie has‖) were excluded from analysis due to their ambiguity in classifying them as social or nonsocial

(explained further below).

An independent observer transcribed approximately 5% of the tapes. ―Matching‖ required both timestamps and vocalizations to be identical between the transcribers. The

Cohen‘s kappa coefficients of reliability for the seven reliability samples ranged from κ =

.38-.79. When the .38 outlier was removed, the median kappa coefficient rose from .67 to .72. Prior to analysis, Cosmo‘s 278 repertoire units were dichotomously categorized as either social or nonsocial. Social vocalizations included those related to physical (e.g.,

―wanna cuddle,‖ ―come here‖) or vocal (e.g., ―wanna whistle,‖ DUW [duet whistle]) interaction, requests beginning with ―wanna‖ or ―Cosmo wanna,‖ greetings (e.g., ―hello,‖

―hi cos‖), farewells (e.g., ―goodbye love you‖), and generic conversation utterances (e.g.

―fine thanks,‖ ―how are you‖). A list of both classes of vocalizations is provided in Table

3.2. An independent observer also categorized the 278 units. The percent agreement between the two observers was 91%.

Procedure

The three sets of transcriptions contained 180 min each of Cosmo alone, talking in the room with BJ, and talking while visually separated from BJ. To assess differences in mean speaking volume, a decibel meter (SkyPaw Co. Ltd, Decibel 10th, v. 3.6.1) was used. To calibrate the decibel meter and the computer (Hewlett-Packard, HP G70) on which the vocalizations were played, a normal speaking volume (60-65 dB, .91 m from decibel meter, Bogardus, Yueh, & Shekelle, 2003) conversation was recorded on the computer‘s built-in recording software (Microsoft Windows 7 Professional, v. 6.1). The computer‘s speaker volume and Windows Media Player volume were adjusted until the decibel meter read approximately 63.0 dB. Once calibrated, the computer speaker volume (88%) and Windows Media Player volume (43%) were maintained throughout all measurements.

A list of 100 randomly sampled time markers from 0-3,300 s (0-55 min) was created with a random number generator. These 100 random points corresponded to a particular time point during each of the taped sessions. The 100 time markers were evenly and randomly distributed among the taped sessions for the three social contexts.

If Cosmo was not vocalizing or if BJ was vocalizing at a particular time point, Cosmo‘s next utterance was recorded. For utterances which were more than one word or sound in length (i.e., units, see Colbert-White et al., 2011), the peak volume in the unit was recorded. In the event a time marker fell during an inappropriate social context (e.g., BJ

infrequently recorded two different 30 min sessions on one tape), a new random time point was generated and used. To assess volume differences between social and nonsocial vocalizations in both conditions, the 100 randomly sampled units in both social conditions were separated into social and nonsocial using the original dichotomous list.

Data Analysis

A preliminary t-test compared the AL and IN conditions‘ mean volumes to determine whether there was an audience effect—that is, to determine if the physical presence of BJ, regardless of her proximity to Cosmo, was associated with a significant change in Cosmo‘s vocalization volume. To establish whether there was a difference in mean vocalization volume for social context and/or vocalization classification, and to assess the interaction between the context and classification, a repeated-measures

ANOVA with two factors was calculated. All statistical analyses were evaluated at the two-tailed, α = .05 level.

Results

Preliminary Analyses

Five AL (Cosmo alone in house) tapes recorded between October 21, 2007 and

December 9, 2007 were included in the analysis. Cosmo‘s volume while alone ranged from 46-89 dB, with a mean volume of 64.65 dB (SD = 11.12). Cosmo‘s vocalizations contained 40 utterance types (i.e., distinct utterances), 5 of which were novel and did not appear in the repertoire established by Colbert-White et al. (2011): ―goodbye love you,‖

―okay back Cosmo,‖ ―okay step up here step up,‖ ―okay wanna go for a walk,‖ and

―peanut in cage.‖ Fifteen nonword sound types, 85 speech utterance types, and 0 speech and nonword combination utterance types were represented. Speech utterances ranged in

volume from 49-68 dB, with a median of 59 dB and a mean of 58.13 dB (SD = 4.57).

Nonword utterances ranged from 46-89 dB, with a median of 70 dB and a mean of 67.57

(SD = 12.03). Sixty-nine of the 100 sampled tokens were nonword sounds, while 31 were speech utterances.

The most commonly uttered vocalizations were NWM [nondescript nonword sound] (20 times), DS [dog bark or growl] (16 times), WBI [wild bird imitation] (8 times), and MWH [nondescript whistle] (6 times). ―We‘re gonna go for a walk,‖ ―wanna peanut,‖ ―that‘s squirrel,‖ OOO [―oooh‖ sound], DUW [duet whistle], and DW [dog whimper or whine] were all uttered 3 times. Only sixteen of the 100 sampled vocalizations were classified as social, precluding the AL condition from subsequent vocalization classification comparisons.

Six IN (Cosmo and BJ in the room) tapes recorded from October 21, 2007 until

January 19, 2008 were included in the analysis. Volumes ranged from 43-88 dB, with a mean volume of 64.66 dB (SD = 10.88). For comparison, 25 of BJ‘s utterances were randomly assessed with the decibel meter. BJ‘s vocalizations ranged from 54-66 dB, with a mean of 63.09 (SD = 7.39), indicating very similar volumes between the two speakers. The vocalizations contained 57 utterance types, 11 of which were novel and did not appear in the repertoire established by Colbert-White et al. (2011): ―be good bird,‖ ―betty jean wanna kiss the feathers,‖ ―cos be a good bird,‖ ―cosmo be a good okay,‖ ―cosmo is a birdie,‖ ―cosmo is a birdie has feathers,‖ ―cosmo‘s a good bird,‖

―don‘t bite okay,‖ ―look wanna kiss KS [kiss sound],‖ ―that‘s tele for bird,‖ and ―what bach.‖ Eight nonword sound types, 88 speech utterance types, and 4 speech and nonword combination utterance types were represented. Twenty-two of the 100 sampled tokens

were nonword sounds, while 78 were speech or speech-nonword combination utterances.

The most commonly uttered vocalizations were DUW (12 times), NWM (6 times), WBI

(6 times), and LS (4 times). Other frequently uttered vocalizations included ―Betty Jean wanna kiss,‖ KS [kiss sound], ―wanna come here,‖ and ―wanna kiss‖ which were all uttered 3 times. Fifty of the vocalizations were classified as social.

Five OUT (Cosmo and BJ in separate rooms out of visual contact) tapes recorded from November 9, 2007 until May 29, 2008 were included in the analysis. Volumes ranged from 50-89 dB, with a mean volume of 72.00 dB (SD = 10.89). The vocalizations contained 45 utterance types, 5 of which were novel and did not appear in the repertoire established by Colbert-White et al. (2011): ―you have reached five hello,‖ ―that‘s clothes,‖ ―LA [BJ laugh mimic] OOO,‖ ―hello hi how are you,‖ and ―a beak.‖ Seventeen nonword sound types, 82 speech utterance types, and 1 speech and nonword combination utterance type were represented. Sixty of the 100 sampled tokens were nonword sounds, while 40 were speech or speech-nonword combination utterances. The most commonly uttered vocalizations were DUW (17 times), NWM (11 times), WBI (9 times), and ―I‘m here‖ (8 times). Other frequently uttered vocalizations included ―here you are,‖ RI

[telephone ring], DO [door open creak], and DW which were all uttered 3 times. Forty- three of the vocalizations were classified as social. In order to conduct the repeated- measures ANOVA, seven randomly sampled social vocalizations were added, and seven randomly sampled nonsocial vocalizations were removed. This resulted in 50 social and

50 nonsocial events. Descriptive data above reflect these alterations. Additional descriptive data are organized in Table 3.3.

Social Condition and Vocalization Classification Volume Differences

The first set of analyses explored the possibility of an audience effect. A t-test revealed no significant difference in overall mean volume between the AL (64.65 dB, SD

= 11.12) and IN (64.66 dB, SD = 10.88) conditions, t(198) = -0.0064, p = .99. However,

Cosmo was significantly louder when BJ was out of the room (64.65 dB, SD = 11.12) compared to when she was completely alone in the house (72.0 dB, SD = 10.90), t(198) =

-4.72, p < .001. These two analyses confirmed that Cosmo‘s volume was unaffected by

BJ‘s physical presence, but was affected by BJ‘s location. The subsequent two-way

ANOVA with social context (IN and OUT) and vocalization classification (social and nonsocial) as factors revealed Cosmo vocalized significantly more loudly during the

2 OUT context than during the IN, F(1, 49) = 18.59, p < .001, ηp = .28.

Student t-tests were conducted to evaluate within condition differences in volume for the IN and OUT conditions (highly uneven numbers of social and nonsocial utterances precluded the AL condition from this and subsequent analyses). Cosmo vocalized social utterances (66.90 dB, SD = 11.45) significantly more loudly than nonsocial utterances (62.42 dB, SD = 9.89) in the IN context, t(98) = 2.09, p = .039.

There was no difference in social (73.88 dB, SD = 11.11) and nonsocial (70.12 dB, SD =

10.45) utterance volumes in the OUT context, t(98) = 1.74, p = .084. The two-way

ANOVA assessed whether Cosmo vocalized utterances classified as social significantly more loudly than nonsocial utterances in the IN and OUT social conditions. The omnibus analysis revealed Cosmo vocalized social utterances significantly more loudly

2 than nonsocial ones, F(1,49) = 10.97, p = .002, ηp = .18.

The final analysis evaluated the predicted interaction effect that Cosmo would vocalize social utterances significantly more loudly during the OUT condition than during the IN condition. The interaction between social context and vocalization classification was not significant, F(1, 49) = .054, p = .82. A visual representation of the main effects, interaction effect, and within context pair-wise effects for the IN and OUT conditions (n = 50 social and nonsocial for both), as well as the AL condition (ns = 16 social, 84 nonsocial) is provided in Figure 3.2.

Discussion

The current study explored volume as a modified quality of vocalizations in the speech and nonword sounds of an African Grey parrot. Two main predictions were tested: 1) Cosmo would vocalize significantly more loudly when her owner was out of the room than when her owner was in the room, and 2) Cosmo would vocalize utterances classified as social significantly more loudly than nonsocial utterances. Though both the

IN and OUT conditions contained many similar vocalizations (e.g., NWM, DUW, WBI), the data supported the first hypothesis; Cosmo did vocalize significantly more loudly in the OUT condition. Additionally, overall, social vocalizations were uttered significantly more loudly than nonsocial vocalizations—supporting the second prediction. However, social utterances were only uttered more loudly than nonsocial in the IN condition. In the

OUT condition, there was no difference in social and nonsocial volumes. Finally, there was no interaction effect between social context and vocalization classification I expected. That is to say, Cosmo did not vocalize social utterances more loudly than nonsocial utterances when BJ was out of the room.

The ability to monitor the social environment and modify qualities of vocalization

(e.g., rate, pitch, volume) occurs in many species. However, in humans, modifying vocalizations to ensure accurate information transfer is termed auditory perspective taking, which falls under the umbrella of theory of mind (Premack & Woodruff, 1978).

Cosmo‘s spontaneous volume modification during times of visual separation adds to the highly debated nonhuman perspective taking literature which suggests that at least some nonhuman species may both monitor and take the cognitive or perceptual perspective of social partners. Such abilities would be highly ecologically relevant to any species where knowing what conspecifics perceive or know is particularly valuable. This is especially true for parrots which live and forage in dense forests; prefer to maintain physical and vocal contact with pair-mates; and have complex fission-fusion social systems, allowing them to benefit from knowing what others‘ know (e.g., Burish et al., 2004; Seibert,

2006).

If Cosmo was not engaging in perspective-taking, three alternative explanations could explain the observed volume differences. First, the increase in volume could be explained by Cosmo‘s body orientation relative to the locations of BJ or the microphone.

To explore this alternative, I reviewed the tapes and found no differences in Cosmo‘s body orientation during the IN and OUT conditions. Additionally, given the location of the microphone (see Figure 3.1), any orienting toward BJ‘s voice would have resulted in higher volume in the IN condition and lower volume in the OUT, which was the opposite of my findings.

The possibility of an audience effect (i.e., the presence of a social partner) explaining the data was also ruled out. Cosmo‘s volume was substantially louder in the

OUT condition compared to the AL condition. This is surprising considering the AL and

OUT conditions shared high percentages of nonword vocalizations (69% and 60%, respectively), which were typically uttered louder than speech sounds. Conversely,

Cosmo‘s volume when alone was no different from when BJ was in the room with her, suggesting her normal ―speaking‖ voice is around 66 dB, very near the human conversation range of 61-65 dB (Bogardus, Yueh, & Shekelle, 2003). Cosmo‘s normal volume increased greatly when BJ was no longer physically present, but still in the home.

As Colbert-White et al. (2011) described, this provides additional evidence that Cosmo is aware both of BJ‘s physical location and willingness to reciprocate interaction. If BJ is willing, but farther away than normal, Cosmo‘s louder vocalizations may indicate basic perceptual perspective-taking to compensate for distance.

Finally, it is possible that Cosmo considers BJ being in a different room to be a distinct vocalization context, rather than an auditory barrier to overcome. According to this mimicry-based hypothesis, Cosmo would vocalize more loudly when BJ is out of the room because BJ herself vocalizes more loudly when she is out of the room. This alternative explanation could be possible considering what BJ yells is likely to be Cosmo- relevant, rather than irrelevant (e.g., phone conversations or self-speech). The lack of an interaction between social context and vocalization classification does support this alternative explanation. However, that Cosmo does not always ―yell‖ learned contact call vocalizations (see Colbert-White et al., 2011) such as ―here you are‖ and ―where are you‖ when she says them in the IN condition provides conflicting evidence for particular vocalizations being learned and vocalized with mimicked volume in a particular setting.

In considering the alternatives, the current findings most likely represent preliminary evidence for perceptual perspective taking, a skill attributed to human theory of mind (Premack & Woodruff, 1978). When BJ was out of the room, Cosmo may have assessed BJ‘s spatial location and responded accordingly with louder vocalizations.

However, not all vocalizations were louder. Many were uttered in her normal voice of around 65 dB while still others were as low as 50 dB and barely registered on the decibel meter. Thus, the secondary analysis evaluated whether the variation in volume was related to whether vocalizations were social or nonsocial in nature. While the analysis found that overall social vocalizations were louder than non social vocalizations, this was not true when BJ was out of the room. This indicates that if perspective taking is occurring, 1) the coding scheme used to classify vocalizations as social and nonsocial may not have been valid, or 2) all vocalizations can serve a social function as long as BJ is home and willing to respond. Vocalizations classified as nonsocial such as ―Betty Jean gonna go in a car‖ or even AM [answering machine beep] could solicit BJ‘s attention, and therefore could still be social.

Further study on this topic is necessary to gain a more holistic view of perspective taking by African Grey parrots. At least one other study has demonstrated visual perceptual perspective taking in this species (Péron et al., 2011). Further, Colbert-White et al. (2011) reported Cosmo was more likely to request physical interaction (e.g., cuddles) than vocal interaction (e.g., talking, whistling) when BJ was verbally ignoring her to talk to ECW. This could be evidence for cognitive perspective taking—that is, BJ is busy and talking to another social partner, so interaction must take a different form.

This provides additional support to my prediction that Cosmo is a strong candidate for

demonstrating perceptual perspective taking. To replicate and extend the current study‘s findings, the procedure could be repeated either with a greater distance separating BJ and

Cosmo while both partners are able to see one another, or from the same distance with a closed door. If Cosmo does, in fact, monitor her surroundings and has a concept of what

BJ can and cannot hear, there may be additional differences in volume. Alternatively, artificial noise like that which was used by Penna et al. (2005) would present a different type of auditory barrier to which Cosmo would have to adapt in order to be heard by BJ.

The current study provided valuable descriptive data on a voice quality other than pitch (e.g., Warren, Patterson, & Pepperberg, 1996) in a captive African Grey parrot.

Cosmo developed a conversational volume that was very similar to her vocal tutor BJ.

Further, her mean vocalization volume was within 1 dB of normal human conversation volume. Without documented volumes for wild African Grey vocalizations, it is impossible to know if Cosmo‘s vocal range is species-typical or if it was learned. In addition to describing the voice of a captive African Grey, I have shown that she modified her vocalization volume during times of visual separation from her owner.

Given features of the species‘ natural history (which includes a dense forest habitat and highly social pair-bonded relationships, Seibert, 2006), it is not surprising that such an ability to monitor surroundings and partners and modify vocalizations accordingly should exist. The findings also suggest that for captive Greys which are pair-bonded to a human caregiver, all vocalizations, regardless of whether they appear (to a human English speaker) as social or nonsocial, are relevant and used to maintain the social relationship.

For captive Greys, then, the content or semantic function of the vocalizations may not always be a priority. Rather, the interaction solicited from the social partner may serve a

particularly salient social feature, no matter whether the vocalizations are beeps, kisses, or ―I love you.‖

Acknowledgements

References

Aeschlimann, M., Knebel, J.-F., Murray, M. M., & Clarke, S. (2008). Emotional pre-

eminence of human vocalizations. Brain Topography, 20, 239-248.

doi:10.1007/s10548-008-0051-8

Aïn, S. Al, Giret, N., Grand, M., Kreutzer, M., & Bovet, D. (2009). The discrimination of

discrete and continuous amounts in African grey parrots (Psittacus erithacus).

Animal cognition, 12, 145-54. doi:10.1007/s10071-008-0178-8

Axia, G. (1996). How to persuade mum to buy a toy. First Language, 16, 301-317. doi:

10.1177/014272379601604803

Bogardus, S. T., Yueh, B., & Shekelle, P. G. (2003). Screening and management of

adult hearing loss in primary care. Journal of the American Medical Association,

289, 1986-1990. doi:10.1001/jama.289.15.1986.

Borke, H. (1975). Piaget's mountains revisited: Changes in the egocentric landscape.

Developmental Psychology, 11(2), 240-244.

Bräuer, J., Call, J., & Tomasello, M. (2004). Visual perspective taking in dogs (Canis

familiaris) in the presence of barriers. Applied Animal Behaviour Science, 88,

299-317. doi:10.1016/j.applanim.2004.03.004

Bucher, T. L. (1983). Parrot eggs, embryos, and nestlings: Patterns and energetic of

growth and development. Physiological Zoology, 56(3), 465-483.

Bugnyar, T., & Heinrich, B. (2005). Ravens, Corvus corax, differentiate between

knowledgeable and ignorant competitors. Proceedings of the Royal Society B,

272, 1641-1646. doi:10.1098/rspb.2005.3144

Burish, M. J., Kueh, H. Y., & Wang, S. S.-H. (2004). Brain architecture and social

complexity in modern and ancient birds. Brain, Behavior and Evolution, 63, 107-

124. doi:10.1159/000075674

Caldwell, D. K., & Caldwell, M. C. (1977). Cetaceans. In: T. A. Sebeok (Ed.), How

animals communicate (pp. 794-808). Bloomington, IN: Indiana University.

Call, J., Bräuer, J., Kaminski, J., & Tomasello, M. (2003). Domestic dogs (Canis

familiaris) are sensitive to the attentional states of humans. Journal of

Comparative Psychology, 117, 257-263. doi:10.1037/0735-7036.117.3.257

Clark, R. A., & Delia, J. G. (1976). The development of functional persuasive skills in

childhood and early adolescence. Child Development, 47, 1008-1014.

doi:10.2307/1128437

Emery, N. J., Seed, A. M., von Bayern, A. M. P., & Clayton, N. S. (2007). Cognitive

adaptations of social bonding in birds. Philosophical Transactions of the Royal

Society B: Biological Sciences, 362, 489-505. doi:10.1098/rstb.2006.1991

Esling, J. H. (1994). Voice quality. In R. Asher, & J. M. Simpson (Eds.). The

encyclopedia of language and linguistics (pp. 4950-4953). Oxford: Pergamon

Press.

Evans, C. S., & Marler, P. (1991). On the use of video images as social stimuli in birds:

Audience effects on alarm calling. Animal Behaviour, 41, 17-26. doi:

10.1016/S0003-3472(05)80499-3

Evans, C. S., & Marler, P. (1994). Food calling and audience effects in male chickens,

Gallus gallus: Their relationships to food availability, courtship and social

facilitation. Animal Behaviour, 47(5), 1159-1170.

Giret, N., Péron, F., Lindová, J., Tichotová, L., Nagle, L., Kreutzer, M., ... & Bovet, D.

(2010). Referential learning of French and Czech labels in African grey parrots

(Psittacus erithacus): Different methods yield contrasting results. Behavioural

processes, 85, 90-98. doi:10.1016/j.beproc.2010.06.010

Hu, Yang, & Cardoso, G. C. (2010). Which birds adjust the frequency of vocalizations in

urban noise? Animal Behaviour, 79, 863-867. doi:10.1016/j.anbehav.2009.12.036

Juola, F. A, & Searcy, W. A. (2011). Vocalizations reveal body condition and are

associated with visual display traits in great frigatebirds (Fregata minor).

Behavioral Ecology and Sociobiology, 65, 2297–2303. doi:10.1007/s00265-011-

1240-0

Kreuz, R., & Roberts, R. (1995). Two cues for verbal irony: Hyperbole and the ironic

tone of voice. Metaphor and Symbolic Activity, 10, 21-30.

doi:10.1207/s15327868ms1001_3

Kundey, S. M. A., De Los Reyes, A., Taglang, C., Allen, R., Molina, S., Royer, E., &

German, R. (2010). Domesticated dogs (Canis familiaris) react to what others can

and cannot hear. Applied Animal Behaviour Science, 126, 45-50.

doi:10.1016/j.applanim.2010.06.002

Kurdek, L. A., & Rodgon, M. M. (1975). Perceptual, cognitive, and affective perspective

taking in kindergarten through sixth-grade children. Developmental Psychology,

11, 643-650. doi:10.1037/0012-1649.11.5.643

Melis, A. P., Call, J., & Tomasello, M. (2006). Chimpanzees (Pan troglodytes) conceal

visual and auditory information from others. Journal of Comparative Psychology,

120, 154-162. doi:10.1037/0735-7036.120.2.154

Michelon, P., & Zacks, J. M. (2006). Two kinds of visual perspective taking.

Perception & Psychophysics, 68(2), 327-337.

Paltridge, B. (2006). Discourse analysis. London: Continuum.

Pasteau, M., Ung, D., Kreutzer, M., & Aubin, T. (2012). Amplitude modulation of sexy

phrases is salient for song attractiveness in female canaries (Serinus canaria).

Animal Cognition, 15, 639-645. doi:10.1007/s10071-012-0492-z

Penna, M., Pottstock, H., & Velasquez, N. (2005). Effect of natural and synthetic noise

on evoked vocal responses in a frog of the temperate austral forest. Animal

Behaviour, 70, 639-651.

Pepperberg, I. M. (1999). The Alex studies: Cognitive and communicative abilities of

Grey parrots. Cambridge, MA: Harvard University Press.

Péron, F., Chadard, C., Nagle, L., & Bovet, D. (2011). Do African grey parrots

(Psittacus erithacus) know what a human experimenter does and does not see?

Behavioural Processes, 87, 237-240. doi:10.1016/j.beproc.2011.04.001

Péron, F., Rat-Fischer, L., Lalot, M., Nagle, L., & Bovet, D. (2011). Cooperative problem

solving in African grey parrots (Psittacus erithacus). Animal Cognition, 14, 545-

553. doi:10.1007/s10071-011-0389-2

Péron, F., Rat-Fischer, L., Nagle, L., & Bovet, D. (2010). ―Unwilling‖ versus

―unable‖: Do grey parrots understand human intentional actions? Interaction

Studies, 11, 428-441. doi:10.1075/is.11.3.06per

Piaget, J., & Inhelder, B. (1963). The child’s conception of space. London: Routledge &

Kegan Paul.

Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind?

Behavioral and Brain Sciences, 1, 515-526. doi:10.1017/S0140525X00076512

Proops, L., & McComb, K. (2010). Attributing attention: The use of human-given cues

by domestic horses (Equus caballus). Animal Cognition, 13, 197-205.

doi:10.1007/s10071-009-0257-5

Santos, L. R., Nissen, A. G., & Ferrugia, J. A. (2006). Rhesus monkeys, Macaca mulatta,

know what others can and cannot hear. Animal Behaviour, 71, 1175-1181.

doi:10.1037/0735-7036.120.2.154

Seibert, L. M. (2006). Social behavior of Psittacine birds. In A. U. Luescher (Ed.),

Manual of parrot behavior (pp. 43-48). Ames: Blackwell Publishing.

Slabbekoorn, H., & Peet, M. (2003). Birds sing at a higher pitch in urban noise. Nature,

424(6946), 267-267.

Stulp, G., Emery, N. J., Verhulst, S., & Clayton, N. S. (2009). Western scrub-jays conceal

auditory information when competitors can hear but cannot see. Biology Letters,

5, 583-585. doi:10.1098/rsbl.2009.0330

Warren, D. K., Patterson, D. K., & Pepperberg, I. M. (1996). Mechanisms of American

English vowel production in a Grey parrot (Psittacus erithacus). The Auk, 113(1),

41-58.

Xitco, M., Gory, J., & Kuczaj, S. (2001). Spontaneous pointing by bottlenose dolphins

(Tursiops truncatus). Animal Cognition, 4, 115-123. doi:10.1007/s100710100107

Table 3.1

Nonword Sound Coding Scheme

CODE DESCRIPTION CODE DESCRIPTION

AM Telephone answering machine beep LS Laser sound

CR Crow caw LSW ―Laser sound-ID-whistle‖ combination

DS Any dog bark, gruff, or howl MWH Miscellaneous one-note whistle

DSS Any dog bark, gruff, or howl sequence NWM Miscellaneous non-whistle

DO Door opening creak NWMS Miscellaneous non-whistle sequence

DOS Door opening creak sequence OOO ―Ooh‖ and other long-o sounds

DUW Duet whistle OU ―Oww‖ as in pain

DUWS Duet whistle sequence OW Owl hoot

DW Dog whine/whimper PH Telephone dialing beep

DWS Dog whine/whimper sequence PHS More than five telephone dialing beeps

FR Frog croak RI Telephone ring

HA Hawk vocalization RIS Telephone ring sequence

HAS Hawk vocalization sequence WBI Wild miscellaneous songbird

ID Indistinguishable WBIS Wild miscellaneous songbird sequence

―Woo-woo-woo‖ or ―woo‖ (B. J.‘s KS Kiss sound WF vocalization of a dog barking)

KSS Kiss sound sequence WW Wolf whistle

LA Laugh WWS Wolf whistle sequence

Note. ―Sequence‖ = Vocalization repeatedly uttered for more than 4 seconds. ―Duet whistle‖ = Melodic whistling, not natural bird vocalization.

Table 3.2

Social and Nonsocial Vocalization Classification

Social Nonsocial

Betty Jean wanna kiss [BJ phone number digits]

Betty Jean wanna kiss the feathers A beak

Come here AM

Come here please Be good bird

Cosmo wanna go for a walk Betty Jean have go in a car

Cosmo wanna go up Come Mary

Cosmo wanna peanut Cos be a good bird

DUW Cos don‘t bite okay

DUW I‘m here Cosmo be a good okay

Fine thank you Cosmo go up

Fine thanks how are you Cosmo go up here

Goodbye Cosmo is a birdie

Goodbye love you Cosmo is a birdie has feathers

Hello Cosmo‘s a good bird

Hello Cos Cosmo we‘re gonna go for a walk

Hello hi how are you CR

Here you are DO

Hi DOS

How are you DS

I love you DSS

I‘m here DW

Kiss DWS

KS Go up

Look Go up here

Look Cosmo Hello Kerri

Look wanna kiss KS Here step up

Okay goodbye Hi Tom

Okay goodbye NWM LA

Wanna LA OOO

Wanna come here LS

Wanna cuddle Mary come on

Wanna kiss MWH

Wanna peanut No

Wanna whi DUW No more peanut

Want kiss NWM

What bach Okay

What‘s bach Okay back Cosmo

What‘s bye Okay Cos

What that Okay go up

What‘s that Okay let‘s go to kitchen

Where are you Okay step up here step up

WW Okay wanna go for a walk

OOO

Peanut

Peanute in cage

Squirrel

Step up

Telephone for bird

That‘s birdie

That‘s clothes

That‘s squirrel

That‘s tele

That‘s tele for bird

Wanna be a good bird

Wanna go back cage

WBI

What a good bird

We‘ll be back soon be back

We‘re gonna go for a walk

We‘re gonna go in a car

You have reached

You have reached five hello

Table 3.3

Volume (dB) of Vocalizations for Classifications and Social Contexts

SOCIAL CONTEXT

AL IN OUT

Social 63.50 (11.02) 66.90 (11.44) 73.88 (11.11)

Nonsocial 64.87 (11.19) 62.42 (9.89) 70.12 (10.45)

Note. Mean decibel recordings for the social vs. nonsocial classification and social context. AL = Cosmo alone in the house (n = 16 social, 84 nonsocial); IN = Cosmo and

BJ in the same room together (n = 50 social, 50 nonsocial); OUT = Cosmo and BJ in adjacent rooms out of visual contact (n = 50 social, 50 nonsocial). Standard deviations appear in parentheses.

Figure 3.1. Illustration of the testing space, including relevant dimensions and furniture.

** ** *

Figure 3.2. Social context (IN the room, OUT of the room, ALone in house) and vocalization classification (social, nonsocial) data.

Both IN and OUT include 50 social and 50 nonsocial vocalizations; AL includes 16 social and 84 nonsocial due to the strong influence of social context on Cosmo‘s vocalizations. Error bars represent standard errors of the means. *p < .05, **p < .001

CHAPTER 4

DISTRIBUTIONS OF SOCIAL- AND NONSOCIAL-THEMED VOCALIZATIONS

FOLLOWING DENIED AND IGNORED REQUESTS BY AN AFRICAN GREY

PARROT1

______

1 Colbert-White, E. N., & Hall, H. C., to be submitted

Abstract

Communicative competence is a measure of individuals‘ sociolinguistic abilities during interactions with social partners. Such abilities are developed through frequent social interactions. The current study aimed to explore communicative competence in a nonhuman speaker, a speech-using African Grey parrot. Spontaneous conversations between one Grey named Cosmo and her caregiver were recorded, from which three corpora were developed: 1) Baseline: No requesting by Cosmo, 2) Ignored Requests:

Excerpts following caregiver ignoring Cosmo‘s requests, and 3) Denied Requests:

Excerpts following caregiver denying Cosmo‘s requests. The distributions of social (e.g.,

―I love you,‖ kiss sounds) and nonsocial (e.g., answering machine beeps, ―That‘s squirrel‖) vocalizations, as well as speech and nonword vocalizations, were nonrandom across the three corpora. Additionally, Cosmo repeated vocalizations when denied and ignored, and interrupted BJ when denied. Neither behavior occurred during the no- request baseline. The data indicated that Cosmo treats being ignored differently from being denied, as evidenced by strategic modification of vocalizations. Such modification represents preliminary evidence for communicative competence in a nonhuman speaker.

Keywords: Communicative competence, African Grey parrot, Psittacus erithacus, language, requesting behavior

Introduction

The extent to which an individual knows what to say and how to say it appropriately in a given social context is a reflection of his or her communicative competence (e.g., Hymes, 1972; Paltridge, 2006). Communicative competence is not simply knowledge of language, but rather knowledge of language and the social dynamics accompanying vocal interactions with social partners. Communicative competence takes into account the social and cultural settings in which an interaction takes place, the nature of the conversation, the relationship between speakers, and societal norms and expectations—thus extending far beyond the grammatical rules of language (Paltridge, 2006).

Like the grammatical features of language, communicative competence is learned through frequent social interaction with others (e.g., Axia, 1996). Cultural differences in behaviors defining communicative competence support the role of learning (e.g.,

Harbaugh, Krause, & Vesterlund, 2007). However, developmental limitations mediate the learning process, especially with respect to higher-order cognitive abilities like perspective taking (e.g., Axia, 1996; Clark & Delia, 1976; Pellegrini, Galda, & Rubin,

1984; Wood, Weinstein, & Parker, 1967). Perspective-taking is argued to require theory of mind—or the ability to understand that another‘s thoughts, feelings, and desires are different from one‘s own (Premack & Woodruff, 1978). Early work established a link between perspective-taking ability and persuasive ability. For example, Clark and Delia

(1976) tested 7-14 year old children and found that as children aged, their use of higher- order persuasive strategies (e.g., supplying advantage to addressee, ―If you let me have a party, I‘ll wash your car for you‖) increased, while use of lower order strategies (e.g., no

support, or demonstration of need for request, ―I would enjoy a party‖) decreased. Very recently, Slaughter, Peterson, and Moore (2013) assessed the relationship among age, persuasive argument generation, and scores on a theory of mind assessment (a false belief test). To assess persuasive argument generation, children aged 3-6 were asked to convince a puppet to eat raw broccoli or brush its teeth. The authors found that as age increased, so did the number of persuasive arguments. When age was controlled for, there was a positive correlation between number of persuasive arguments and scores on the theory of mind assessment, linking both age and theory of mind to competency in persuasion.

Along with persuasive strategies, requesting behavior is one of many social strategies under the umbrella of communicative competence that must be mastered early on. Requesting behavior can be classified as an interactive control act where the purpose is to attract the attention of a social partner and make a persuasive case that convinces the partner to act, while still maintaining the social relationship (Stein & Albro, 2001; Ervin-

Tripp, 1982). Age-defined qualities of the initial request (e.g., politeness) and the persuasiveness of the request (or lack thereof) all contribute to whether a request is granted.

The earliest requesting behavior, gesturing by pre-linguistic infants (between 1 and 2 years), incorporates pointing, vocalizing, and visual checking with the social partner. In response to being ignored, children as young as 1.5 have a concept that being ignored is different from being refused and respond by differentially repeating or reformulating requests (Marcos & Bernicot, 1994, 1997). As children develop, perspective-taking, additional social and cognitive skills, and language become mediating

factors in the effectiveness of requesting behavior (Stein & Albro, 2001). For the youngest children, those who have not yet developed the ability to take the perspective of others, speech acts associated with forming an argument/request and responding to denied or ignored requests are much different from those of older children and adults

(e.g., Clark & Delia, 1976; Ervin-Tripp, Guo, & Lampert, 1990; Wootton, 1981). While children under age 3 modify request form (e.g., politeness) based on the identity of the addressee (Ervin-Tripp et al., 1990), their request tactics are basic and offer little support for those requests (i.e., the ―why‖ behind the request, Clark & Delia, 1976). Further, these youngest children respond to unmet requests more often with ―insistence strategies‖ as Stein and Albro (2001) call them: repetition, repetition with aggravation, threatening, or pleading (e.g., Clark & Delia, 1976; Ervin-Tripp et al., 1990; Garvey, 1975), and are considered incompetent at recovering from refusal (Garvey, 1975).

By 4 years old, children still rely heavily upon repetition of requests (Wootton,

1981), but can provide more justification (Clark & Delia, 1976; Ervin-Tripp et al., 1990), suppress frustration after refusals (Ervin-Tripp et al., 1990), repeat the addressee‘s name before retrying an ignored request (Garvey, 1975), and modify their argument strategies in response to the social partner‘s identity (Albro, 2001). These are all indicators that by this age, some children may be able to take the perceptual perspective of others into account (e.g., ―Maybe mom couldn‘t hear me‖).

By age 5 or 6 (depending upon the task provided), children will consider the desires of others, reflecting the clear presence of perspective taking (Axia, 1996; Clark &

Delia, 1976). Around this same age, children‘s (> 6 years) persuasive behavior incorporates compromise and bargaining (Weiss & Sachs, 1991). For example, Weiss

and Sachs (1991) refused 3-6 year old children‘s persuasive attempts during a lab-based, role play task and noted that bargaining behavior (e.g., ―I‘ll give you a million dollars if...‖) emerged by age 6 (see Bartsch, Wright, & Estes, 2010, for evidence of bargaining in everyday conversation by younger children). It is not until adolescence that the greatest attunement of the persuader to social partners‘ psychological states occurs (Clark

& Delia, 1976). Thus, while much of requesting and denied request behavior is learned, predictable patterns do surface as sociocognitive abilities develop.

Like humans, other social species must maintain positive relationships during interactions with conspecifics. It is not surprising, then, that parallels to human requesting strategies can be found in the wild. For example, some nonhuman primate species have ―politeness‖ gestures accompanying behavioral requests (e.g., chimpanzees,

Pan troglodytes, present an up-turned palm, Corballis, 2003). Evidence for strategic use of persistence and strategic elaboration of gestures in wild chimpanzees has also been found (e.g., Leavens, Russell, & Hopkins, 2005; Roberts, Vick, & Buchanan-Smith,

2013). Roberts et al. (2013) reported that when a conspecific‘s response partially matched the sender‘s goal, the sender repeated the gesture; however, when the receiver‘s action did not match the intended behavior at all, the sender signaled a different gesture with a similar function.

Examples of requesting politeness in nonhumans‘ vocal/symbolic communication are limited. Lana the chimpanzee was trained to use Yerkish, a symbol-based artificial language (Rumbaugh, 1977). Lana was documented using the symbol for ―please‖ prior to requesting objects like music or water. It is important to note, that the ―please‖ symbol started the keyboard, and so was necessary to begin communicative interaction. One of

Fouts‘ (1997) chimpanzees, Washoe, was trained to use ―please‖ in the form of a manual gesture. Fouts describes one signed conversation between Washoe and a volunteer who had just had a miscarriage. The volunteer signed to Washoe MY BABY DIED. In response, Washoe held the volunteer, and signed PLEASE PERSON HUG. At the surface, this use of ―please‖ more closely approximates humans‘ use of the word than

Lana‘s required use of the ―please‖ symbol to begin requests. Aside from these lab examples, to this author‘s knowledge, there exists no literature on learned requesting behavior in nonhumans, and none on repair or strategies following a denied request.

The current study explored one speech-using African Grey parrot‘s requesting behavior and responses to denied and ignored requests during conversations with her female caregiver, Betty Jean. Conversations were compared qualitatively and statistically during three manipulated speech contexts: Baseline (no requests), Ignored Requests, and

Denied Requests. We predicted that Cosmo the parrot‘s verbal behavior would be similar to that of children in the early stages of cognitive and linguistic development.

Specifically, we hypothesized that Cosmo would respond differentially to being ignored and denied. Additionally, we hypothesized that Cosmo‘s most frequent response to being ignored or denied requests would be repetition (similar to children under age 4, Clark &

Delia, 1976; Ervin-Tripp et al., 1990; Garvey, 1975).

Recent evidence suggests that African Grey parrots may engage in perceptual perspective taking (Péron, Chadard, Nagle, & Bovet, 2011; see also Chapter 3 of this dissertation). Additionally, Pepperberg (1999) has compared some features of Greys‘ cognitive behavior to that of 6-year-old children. For these reasons, evidence for more

advanced requesting strategies such as bargaining was considered possible, but was not hypothesized.

The hypotheses we developed assumed that requesting behavior (gestural or vocal), regardless of species, is associative in nature. That is to say, while higher-order cognitive abilities surely mediates features of requesting behavior, individuals learn through trial and error with old and new social partners what ―works‖ and what does not in formulating and navigating requests. Approaching the study of Cosmo‘s requesting and persuasive strategies through an associative framework allows for objective interpretations to be made which do not assume intentionality of vocalizations.

Method

Subject and Housing

Cosmo was 6 years old at the beginning of the project in 2007. Cosmo‘s female owner (BJ) purchased Cosmo from a pet store in 2002 when she was 5 months old.

Although Cosmo had some experience hearing other human speakers, BJ was her consistent companion. BJ established a simplified grammar with limited vocabulary by labeling new objects for Cosmo and correcting Cosmo‘s misuse and mispronunciation of words. Social interaction with BJ was the primary means by which Cosmo acquired melodies, English speech, and some nonword sounds like kiss noises. Additional nonword sounds such as microwave beeps were acquired naturally. Two female dogs also resided with BJ and Cosmo at the time of data collection.

All video-taping was conducted at BJ‘s home with Cosmo in her primary cage

(55.9 x 61.0 x 83.8 cm, with perch extending 40.6 cm from top) which was located in a

sun room facing BJ‘s reading chair 2 m away. Food and water were provided ad libitum throughout all testing.

Recording and Transcriptions

For three weeks prior to testing, Cosmo was habituated to a Sony DCR-TRV39 mini-DV video camera on a tripod 1.5 m from the cage. Cosmo‘s cage was the only object in the camera frame. The camera‘s built-in microphone (32 kHz, 16-bit audio) recorded all audio. After the habituation phase, BJ recorded five, hour-long events in each of two social conditions during times BJ thought Cosmo would be particularly talkative. In the IN condition, BJ sat in her reading chair and interacted with Cosmo as normal in the room with her. In the OUT condition, BJ remained out of sight in an adjacent room, but interacted with Cosmo as normal. An illustration of the layout of the space is shown in Figure 4.1.

Videos were transcribed using the code ―ID‖ (i.e., indistinguishable) whenever vocalizations were not clear. Syllables and fragments were transcribed as they were heard (e.g., ―tele‖ and ―showe‖ as in ―telephone‖ and ―shower,‖ respectively). To decrease experimenter bias, contextual information in the videos was not used to construct the transcriptions. Nonword sounds were transcribed with two- or three-letter codes (refer to Table 4.1 for nonword sound coding scheme). While Colbert-White et al.

(2011) excluded units which were uttered only once, due to the small size of the dataset, all units were included in analysis. Fragments (e.g., ―Mary is a doggie has‖) were excluded from analysis due to their ambiguity in classifying them as social or nonsocial

(explained further below).

An independent observer transcribed approximately 5% of the tapes. ―Matching‖ required both timestamps and vocalizations to be identical between the transcribers. The

Cohen‘s kappa coefficients of reliability for the seven reliability samples ranged from κ =

―wanna cuddle,‖ ―come here‖) or vocal (e.g., ―wanna whistle,‖ DUW [duet whistle]) interaction, requests beginning with ―wanna‖ or ―Cosmo wanna,‖ greetings (e.g., ―hello,‖

―hi cos‖), farewells (e.g., ―goodbye love you‖), and generic conversation utterances (e.g.

―fine thanks,‖ ―how are you‖). A list of both classes of vocalizations is provided in Table

3.2. An independent observer also categorized the 278 units. The percent agreement between the two observers was 91%.

The SAE Phrase Frequency Tool (Strategic Analysis Enterprises, Inc.,

Williamsburg, VA) computer program searched for recurrent phrases, and not only single words. This program is used by linguists to find recurrent phrases in text. The program tabulated the number of occurrences of all words and phrases ranging from one to nine words long (nonword sounds were considered words) that occurred at least once. To aid with qualitative analysis of the conversations, utterances that only occurred once were included, despite the inability to make comparisons across conditions. If a phrase only occurred as part of a larger phrase (e.g., ―in a car‖ was only uttered as ―go in a car‖), then the program tabulated the larger of the two phrases. Phrases were not included if they contained or spanned periods.

100

Procedure

One-hundred eighty minutes both of IN and OUT footage were used in the analyses. Transcriptions were prepared for analysis using AntConc (v. 3.2.1w) freeware text analysis program. Keyword searches identified all of Cosmo‘s requests, from which the four most frequently requested items (cuddles, approaching Cosmo‘s cage, being let out of cage, and peanuts) were included in all analyses. The four types of requests were first divided into whether the request was denied (i.e., direct refusal) or ignored (i.e., no response from BJ within 30 s, or ignored but redirected by BJ such as ―Let‘s play telephone‖). Cosmo‘s vocalizations during the 2 min following refused and ignored requests were extracted. A third corpus of 44 min without any requests from 7 IN and 1

OUT transcriptions served as the Baseline condition for comparison.

Data Analysis

Comparisons among Ignored Requests, Denied Requests, and Baseline vocalizations were made first using a series of three Spearman‘s correlations (Baseline-

Ignored, Baseline-Denied, Ignored-Denied) for evidence of differential use of individual vocalizations. Additionally, comparisons of distributions of word and nonword units, and social and nonsocial units were made among the three corpora using chi-square tests. All statistical analyses were evaluated at α = .05 level.

Results

Preliminary Findings

After combining vocalizations using the SAE phrase frequency table and

AntConc, the Ignored Request corpus contained 54 unit types and 111 tokens. The corpus consisted of 38 speech units, 11 nonword sounds, and 5 speech-nonword sound

101

combinations. Sixty-eight vocalizations were social; 43 were nonsocial. The five most frequently uttered vocalizations were DUW ([Duet whistle], 16 times), KS ([Kiss sound],

9 times), ―No‖ (7 times), NWM ([Miscellaneous nonword sound] 5 times), and ―I‘m here‖ (5 times). Nonword sounds comprised 40% of the tokens. Vocalizations were predominantly requests for physical and vocal interaction (e.g., Betty I wanna kiss on the beak KS KS, Betty wanna kiss feathers MWH, Come here, Come here Cosmo wanna cuddle, Cosmo wanna come here, Cosmo wanna go to Betty, Cosmo wanna talk, WW

[Wolf whistle]), repeated requests and statements about getting out of the cage (e.g.,

Okay go up, Wanna go up, Cosmo go up, Cosmo wanna go up), as well as five other requests beginning with ―wanna‖ (Wanna be a good bird, Wanna come here, Wanna cuddle, Wanna go bed, Wanna stay here).

The Denied Request corpus contained 139 unit types and 334 tokens. The corpus consisted of 112 speech units, 15 nonword sounds, and 12 speech-nonword sound combinations. One-hundred and fifty-two vocalizations were social; 182 were nonsocial.

The five most frequently uttered vocalizations were DUW (33 times), NWM (25 times),

―No‖ (14 times), KS (12 times), and OOO ([―oooh‖ sound], 11 times). Nonword sounds comprised 34% of the tokens. No clear patterns emerged in the content of the vocalizations, with the exception of a presence of utterances relevant to BJ‘s dogs (e.g.,

Hello Kerri, Mary, Mary has feathers MWH, Mary‘s a dog), which was not in the

Ignored and Baseline corpora. Requests and statements relevant to BJ kissing, whistling, or approaching Cosmo‘s cage, and repetition of requests and statements relevant to being let out of the cage were also present.

102

The Baseline corpus contained 91 unit types (i.e., distinct utterance units) and 272 tokens (i.e., total utterances). The corpus consisted of 69 speech unit types, 16 nonword sounds, and 6 speech-nonword sound combinations (e.g., ―DUW I‘m here‖). One- hundred and four vocalizations were social; 168 were nonsocial. The five most frequently uttered vocalizations were NWM (26 times), DUW (25 times), WBI ([Wild bird imitation], 22 times), PH ([Telephone beep], 16 times), and WW (11 times).

Nonword sounds comprised 51% of the tokens. Vocalizations were predominantly statements beginning with Cosmo (e.g., Cosmo be a good bird please, Cosmo poop,

Cosmo has feathers MWH, Cosmo‘s a bird, Cosmo‘s a girl, Cosmo‘s good bird), greetings and farewells, and labels (e.g., That‘s bark, That‘s Betty kiss, That‘s birdie,

That‘s Cosmo, That‘s tele).

Quantitative Comparisons

The 20 most frequently uttered vocalizations for each of the three corpora appear in Table 4.3. Three Spearman‘s correlations determined there was no significant relationship among the rankings of the 20 most frequently uttered vocalizations for the three corpora pairs: Baseline-Denied, rs = 0.35, Denied-Ignored, rs = -0.056, Baseline-

Ignored, rs = 0.039, all ps > .05. The series of correlations supported qualitative observations that the content of Cosmo‘s vocalizations was significantly different in post- ignored vocalizations, post-denial vocalizations, and non-request vocalizations.

To assess whether the distribution of speech and nonword vocalizations was nonrandom for the three corpora, a chi-square test was conducted. The analysis showed that the rates with which Cosmo uttered speech and nonword vocalizations was significantly different across the three corpora X2(2, N = 690) = 16.71, p = .00024. That

103

is, Cosmo uttered speech and nonword vocalizations with differing rates immediately following an ignored request (61 speech, 44 nonword), immediately following a denied request (207 speech 112, nonword), and when she was not requesting anything of BJ (128 speech, 138 nonword). These data are summarized as proportions in Figure 4.2.

A second chi-square test assessed whether the distribution of social and nonsocial vocalizations was nonrandom for the three corpora. The analysis revealed that the rates with which Cosmo uttered social and nonsocial vocalizations was different across the three corpora, X2(2, N = 718) = 16.91, p = .00021. That is to say, Cosmo uttered social and nonsocial vocalizations with differing rates immediately following an ignored request

(68 social, 43 nonsocial), immediately following a denied request (153 social, 182 nonsocial), and when she was not requesting anything of BJ (104 social, 168 nonsocial).

These data are summarized as proportions in Figure 4.3.

Qualitative Observations

Differences in rates of vocalization repetition were observed across the three corpora. There was no evidence of repetition of vocalizations in the baseline corpus.

However, Cosmo did frequently repeat her requests when she was ignored or denied.

This occurred regardless of whether BJ was in the room or out of the room, indicating more that the purpose of repetition was not to ensure that BJ heard her (refer to Chapter 3 for evidence of auditory perceptual perspective taking). As shown in transcription excerpts in Table 4.4, repetition following unmet requests was sometimes accompanied by solicitations for interaction or possible redirection. For example, in Excerpt A, following three identically repeated and refused requests to be let out of her cage, Cosmo solicited interaction from BJ, then returned to asking to be let out. Excerpt B presents a

104

different type of repetition from Excerpt A—alternative form repetition—in which

Cosmo vocalized multiple utterances synonymous with the theme of being let out of her cage. Similar to Excerpt A, Cosmo responded to one minute of being ignored by BJ by saying ―hi‖ (which elicited a response from BJ), then repeating her request after BJ replied.

As demonstrated in Chapter 2, Cosmo and BJ maintained turn-taking in their conversations. However, there were differences between the baseline and the denied and ignored request corpora. In particular, the baseline contained four instances of simultaneous speech, where BJ and Cosmo both tried to take up the turn at the same time.

Usually this occurred following a silence or during sessions when BJ was out of the room. For example, Cosmo asked ―What that?‖ then proceeded to say ―What‖ at the same time as BJ laughed. There were also two instances of interruption by BJ while

Cosmo was vocalizing. In one case, Cosmo was vocalizing a long wild bird imitation, and BJ said ―Cosmo,‖ possibly out of frustration given the tone of voice. In the other interruption, BJ said ―Wow‖ over a long nonwhistle nonword sound that Cosmo made.

There was no evidence of Cosmo interrupting BJ in the baseline; however, there were two instances of Cosmo interrupting BJ during unmet requests. As shown in Table 4.5, following a series of refusals, Cosmo interrupted BJ, causing BJ to stop vocalizing, in order to repeat her requests to be let out of her cage.

During three of the recording sessions, BJ did let Cosmo out. This violated the instruction we gave BJ to refuse Cosmo‘s requests to be let out; however, it did provide an opportunity to observe Cosmo‘s vocal patterns during initially unmet requests which were later met. During all of these special cases, once Cosmo was let out of her cage, she

105

stopped requesting to be let out. The example in Table 4.6 shows Cosmo was very persistent in her requests, despite being ignored and refused. Once she was out, she abandoned all cage requests in favor of kiss requests. Along a similar vein, Cosmo also maintained her requests in a goal-oriented fashion. The examples in Table 4.7 demonstrate how BJ‘s attempts to redirect Cosmo were rarely effective at making Cosmo stop repeating requests (Excerpt A), and even when Cosmo redirected herself via a new request, she still returned to the original request (Excerpt B).

To investigate strategic use of vocalizations during requesting, I recorded patterns in Cosmo‘s use of two phrases, ―Wanna be a good bird‖ and ―Cosmo don‘t bite okay.‖

These frequently accompanied Cosmo‘s requests to be let out of her cage (see Table 4.8).

These phrases were most likely said by BJ during times when Cosmo asked to be let out of her cage. Neither of these phrases appeared during the baseline, which confirmed they were specific and appropriate to the context of letting Cosmo out of her cage. Despite the phrases being associated with being let out of the cage, they were not uttered when BJ met Cosmo‘s request and let her out of her cage, despite the fact that BJ most likely would utter the phrases to Cosmo once she was let out (see Table 4.6 for example).

Discussion

The current study examined conversations between an African Grey parrot and her female caregiver for evidence of communicative competence. The recorded conversations included spontaneous instances of both denied and ignored requests to establish how the parrot modified her vocalizations in response to relationship denied and ignored requests. We approached the analyses from an associative perspective. Just as young children learn through experience what ―works‖ and what does not ―work‖ when

106

making requests and responding to unmet requests, we hypothesized that an African Grey parrot might also demonstrate these skills given Greys‘ renowned cognitive and linguistic capabilities.

The results with Cosmo mirror that of young children‘s conversations. By age 4, children develop the ability to treat an ignored request as a failed attempt, and modify by repeating the social partner‘s name before repeating the request; whereas, directly refused requests are treated differently (Ervin-Tripp et al., 1990; Garvey, 1975; Wootton, 1981; see Marcos & Bernicot, 1994, 1997 for evidence of differential response to being ignored and refused in children under 2 years). Cosmo, too, treats being ignored and denied differently. Previous work with Cosmo reported that during times when BJ ignored

Cosmo in favor of talking to ECW, Cosmo‘s vocalization rate decreased substantially to only a few utterances per hour-long session (Colbert-White et al., 2011). Additionally, being vocally ignored was associated with a proportional increase in requests for physical interaction, a finding that Colbert-White et al. interpreted as Cosmo modifying her requests in response to BJ‘s reactions. This previous study and current results both support the notion that Cosmo not only notices when she is being ignored, but vocalizes differentially to achieve her requesting goals.

The results supported the two main hypotheses. Cosmo‘s word choice following being denied and ignored was significantly different not only from the no-request baseline, but also from each other. Further, qualitative observations showed differential patterns in repetition, interruption, and placement of utterances within conversation.

Taken together, Cosmo did modify her vocalizations to BJ based on context provided by

107

the speech environment, which is Hymes‘ (1972) definition of communicative competence.

Only Pepperberg (1988) has used the term communicative competence to describe nonhumans‘ vocal behavior. Specifically, Pepperberg described the similarly influential role of social interaction in the development of communicative competence by humans and avians which learn their vocalizations from adult conspecifics. The more structured definition of communicative competence (i.e., knowing what to say and how to say it given a social context) was not addressed by Pepperberg. Nevertheless, the results of this study have shown that Cosmo‘s strategic use of vocalizations appears to fit within the human language definition of communicative competence, which opens the door to other social species‘ having the cognitive wherewithal to demonstrate this ability.

Communicative competence implies intention and theory of mind in humans; researchers advocating the existence of theory of mind in nonhumans may benefit from exploring the linguistics literature to develop measures to test their claims.

Alternatively, rather than presenting evidence for communicative competence, the current study‘s findings could be interpreted as social context providing a cue for how to respond. For example, following denied requests to get out of her cage, Cosmo uttered more vocalizations associated with the social context of being let out of her cage. This included mimics of BJ telling Cosmo to be good and not to bite. Uttering these vocalizations during the context of Cosmo asking to be let out of her cage could represent a contextually-learned behavior (i.e., vocal behavior). During the social and vocal context of being let out of the cage, Cosmo could be uttering all vocalizations surrounding that setting. This alternative explanation, however, does not appear to

108

explain the findings. Evidence is found in the above example. The word ―bite‖ was uttered 1 time when ignored compared to 16 times when denied (a frequency distribution which would be predicted by chance only 7% of the time), and once Cosmo was let out, she stopped using words like ―bite,‖ though BJ would often remind her not to once she was let out. Rather than mimicking vocalizations surrounding a static social context, the results indicate that Cosmo instead first proposes a request, and then modifies subsequent vocalizations in accordance with each new social circumstance. That is to say, there is strategic use of vocalizations which are dependent upon the social partner‘s actions, not a particular event (e.g., getting out of the cage).

Cosmo‘s differential responding to two similar but distinct types of requests

(ignored vs. denied) reflects how finely-tuned she is in her monitoring and responding to

BJ. Péron et al.‘s (2010) work with African Greys corroborated these findings from a behavioral perspective in their finding that Greys behaved differently when a social partner was unable versus unwilling to provide treats. These authors, like others (e.g.,

Premack & Woodruff, 1978), explored theory of mind as an explanation for an understanding of what a social partner can and cannot do. Statistical differences in

Cosmo‘s rate of use of different types of vocalizations, as well as qualitative differences in vocal behaviors like repetition and interruption following unmet requests support the hypothesis that African Greys may have the cognitive wherewithal to understand and respond to the intentions and knowledge of others. This finding supports previous nonhuman theory of mind work, and provides the first empirical study arguing for communicative competence from a linguistic perspective in an African Grey parrot.

109

In the future, probabilistic models of Cosmo‘s conversations with BJ may allow for more objective measures of intentionality. Without being able to ask Cosmo the intention behind her placement of phrases like ―Cosmo don‘t bite okay‖ following a refusal to be let out of her cage, it is inappropriate to conclude with confidence that any observed patterns represent intentional placement. However, given the 278 vocalizations in Cosmo‘s repertoire (Colbert-White, Covington, & Fragaszy, 2011), the probability of particular vocalizations occurring, and then occurring following a particular response from BJ, can be assessed. For example, Ervin-Tripp et al. (1990) noted that mitigating words like please after refusals appear in greater frequencies in the requesting events of

3.5-year olds, suggesting strategic placement in response to the new speech context of refusal. Developing such a model would strengthen claims about Cosmo‘s use of her repertoire. In particular, intentionality, manipulation, and perhaps behavioral indicators of frustration (e.g., interruption, Gallager & Craig, 1982) could be assessed in more detail.

The current study showed through a variety of measures that one captive African

Grey has learned to use an artificial communication system in a goal-oriented manner during instances of relationship conflict brought about by denied or ignored requests.

These results provide evidence for strategic requesting behavior by a Grey parrot. Such behavior has already been noted in other nonhumans, as with wild chimpanzees‘ modification of communicative gestures based on the receivers‘ response (e.g., Leavens et al., 2005; Roberts et al., 2013). The results with Cosmo are of particular interest because the vocalizations Cosmo uses to communicate with BJ are not species-specific.

Rather, due to her natural history as a social, vocal learner, Cosmo acquired and

110

effectively uses BJ‘s communication system. Whether or not wild African Greys monitor and modify species-typic vocalizations similar to what has been demonstrated with wild chimpanzees has yet to be determined. However, Cosmo‘s ability to modify her species atypical communication system may be indirect evidence that wild Greys may also modify their vocalizations, a valuable finding for those studying animal communication systems.

111

Acknowledgements

Marisol Macias aided with literature searching. In addition, we thank Allison

Kaufman for assisting with reliability coding and associated calculations and Dominic

Byrd for his illustration of the testing area in Figure 4.1.

112

References

Axia, G. (1996). How to persuade mum to buy a toy. First Language, 16, 301-317. doi:

10.1177/014272379601604803

Bartsch, K., Wright, J. C., & Estes, D. (2010). Young children‘s persuasion in everyday

conversation: Tactics and attunement to others‘ mental states. Social

Development, 19, 394-416. doi:10.1111/j.1467-9507.2009.00537.x

Clark, R. A., & Delia, J. G. (1976). The development of functional persuasive skills in

childhood and early adolescence. Child Development, 47, 1008-1014.

doi:10.2307/1128437

Corballis, M. C. (2003). From hand to mouth: The origins of language. Princeton, NJ:

Princeton University Press.

Ervin-Tripp, S. M., (1982). Ask and it shall be given you: Children's requests. In: H.

Byrnes, (ed.), Georgetown roundtable on languages and linguistics, vol. 35 (pp.

235-245). Washington, DC: Georgetown University.

Ervin-Tripp, S., Guo, J., & Lampert, M. (1990). Politeness and persuasion in children‘s

control acts. Journal of Pragmatics, 14(2), 307-331.

Fouts, R. (1997). Next of kin: My conversations with chimpanzees. New York:

HarperCollins Publishers, Inc.

Gallagher, T. M., & Craig, H. K. (1982). An investigation of overlap in children‘s

speech. Journal of Psycholinguistic Research, 11, 63-75.

doi:10.1007/BF01067502

Garvey, C. (1975). Requests and responses in children‘s speech. Jouranl of Child

Language, 2, 41-63. doi: 10.1017/S030500090000088X

113

Harbaugh, W. T., Krause, K., & Vesterlund, L. (2007). Learning to bargain. Journal of

Economic Psychology, 28, 127-142. doi:0.1016/j.joep.2006.03.005

Hymes, D. (1972). On communicative competence. In J. B. Pride & J. Holmes (Eds.),

Sociolinguistics: Selected readings (pp. 269-293). Harmondsworth: Penguin.

Leavens, D. A., Russell, J. L., & Hopkins, W. D. (2005). Intentionality as measured in

the persistence and elaboration of communication by chimpanzees (Pan

troglodytes). Child Development, 76, 291-306. doi: 10.1111/j.1467-

8624.2005.00845.x

Marcos, H., & Bernicot, J. (1994). Addressee co-operation and request reformulation in

young children. Journal of Child Language, 21, 677-692.

doi:10.1017/S0305000900009508

Marcos, H., & Bernicot, J. (1994). How do young children reformulate assertions? A

comparison with requests. Journal of Pragmatics, 27, 781-798.

doi:10.1016/S0378-2166(96)00062-8

Paltridge, B. (2006). Discourse analysis. London: Continuum.

Pellegrini, A. D., Galda, L., & Rubin, D. (1984). Persuasion as a social-cognitive activity:

The effects of age and channel of communication on children‘s production of

persuasive messages. Language & Communication, 4, 285-293.

doi:10.1016/0271-5309(84)90012-0

Pepperberg, I. M. (1988). The importance of social interaction and observation in the

acquisition of communicative competence: Possible parallels between avian and

human learning. In T. R. Zentall, & B. G. Galef (Eds.), Social learning:

114

Psychological and biological perspectives, (pp. 279-301). Hillsdale, NJ:

Lawrence Erlbaum Associates.

Pepperberg, I. M. (1999). The Alex studies: Cognitive and communicative abilities of

Grey parrots. Cambridge, MA: Harvard University Press.

Péron, F., Chadard, C., Nagle, L., & Bovet, D. (2011). Do African grey parrots

(Psittacus erithacus) know what a human experimenter does and does not see?

Behavioural Processes, 87, 237-240. doi:10.1016/j.beproc.2011.04.001

Péron, F., Rat-Fischer, L., Nagle, L., & Bovet, D. (2010). ‗Unwilling‘ versus ‗unable‘:

Do grey parrots understand human intentionally action? Interaction Studies, 11,

428-441. doi:10.1075/is.11.3.06per

Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind?

Behavioral and Brain Sciences, 1, 515-526. doi:10.1017/S0140525X00076512

Rumbaugh, D. M. (1977). Language learning by a chimpanzee: The Lana project. New

York: Academic Press.

Slaughter, V., Peterson, C. C., & Moore, C. (2013). I can talk you into it: Theory of mind

and persuasion behavior in young children. Developmental Psychology, 49, 227-

231. doi:10.1037/a0028280

Stein, N. L., & Albro, E. R. (2001). The origins and nature of arguments: Studies in

conflict understanding, emotion, and negotiation. Discourse Processes, 32, 113-

133. doi:10.1080/0163853X.2001.9651594

Weiss, D. M., & Sachs, J. (1981). Persuasive strategies used by preschool children.

Discourse Processes, 14, 55-72. doi: 10.1080/01638539109544774

115

Wood, J. R., Weinstein, E. A., & Parker, R. (1967). Children's interpersonal tactics.

Sociological Inquiry, 37, 129-138. doi: 10.1111/j.1475-682X.1967.tb00644.x

Wootton, A. J. (1981). Two request forms of four year olds. Journal of Pragmatics, 5(6),

511-523.

116

Table 4.1

Nonword Sound Coding Scheme

CODE DESCRIPTION CODE DESCRIPTION

AM Telephone answering machine beep LS Laser sound

CR Crow caw LSW ―Laser sound-ID-whistle‖ combination

DS Any dog bark, gruff, or howl MWH Miscellaneous one-note whistle

DSS Any dog bark, gruff, or howl sequence NWM Miscellaneous non-whistle

DO Door opening creak NWMS Miscellaneous non-whistle sequence

DOS Door opening creak sequence OOO ―Ooh‖ and other long-o sounds

DUW Duet whistle OU ―Oww‖ as in pain

DUWS Duet whistle sequence OW Owl hoot

DW Dog whine/whimper PH Telephone dialing beep

DWS Dog whine/whimper sequence PHS More than five telephone dialing beeps

FR Frog croak RI Telephone ring

HA Hawk vocalization RIS Telephone ring sequence

117

HAS Hawk vocalization sequence WBI Wild miscellaneous songbird

ID Indistinguishable WBIS Wild miscellaneous songbird sequence

―Woo-woo-woo‖ or ―woo‖ (B. J.‘s KS Kiss sound WF vocalization of a dog barking)

KSS Kiss sound sequence WW Wolf whistle

LA Laugh WWS Wolf whistle sequence

Note: ―Sequence‖ = Vocalization repeatedly uttered for more than 4 seconds. ―Duet whistle‖ = Melodic whistling, not natural bird vocalization.

118

Table 4.2

Social and Nonsocial Vocalization Classification

Social Nonsocial

Betty Jean wanna kiss [BJ phone number digits]

Betty Jean wanna kiss the feathers A beak

Come here AM

Come here please Be good bird

Cosmo wanna go for a walk Betty Jean have go in a car

Cosmo wanna go up Come Mary

Cosmo wanna peanut Cos be a good bird

DUW Cos don‘t bite okay

DUW I‘m here Cosmo be a good okay

Fine thank you Cosmo go up

Fine thanks how are you Cosmo go up here

Goodbye Cosmo is a birdie

Goodbye love you Cosmo is a birdie has feathers

Hello Cosmo‘s a good bird

Hello Cos Cosmo we‘re gonna go for a walk

Hello hi how are you CR

Here you are DO

Hi DOS

How are you DS

I love you DSS

119

I‘m here DW

Kiss DWS

KS Go up

Look Go up here

Look Cosmo Hello Kerri

Look wanna kiss KS Here step up

Okay goodbye Hi Tom

Okay goodbye NWM LA

Wanna LA OOO

Wanna come here LS

Wanna cuddle Mary come on

Wanna kiss MWH

Wanna peanut No

Wanna whi DUW No more peanut

Want kiss NWM

What bach Okay

What‘s bach Okay back Cosmo

What‘s bye Okay Cos

What that Okay go up

What‘s that Okay let‘s go to kitchen

Where are you Okay step up here step up

WW Okay wanna go for a walk

OOO

120

Peanut

Peanute in cage

Squirrel

Step up

Telephone for bird

That‘s birdie

That‘s clothes

That‘s squirrel

That‘s tele

That‘s tele for bird

Wanna be a good bird

Wanna go back cage

WBI

What a good bird

We‘ll be back soon be back

We‘re gonna go for a walk

We‘re gonna go in a car

You have reached

You have reached five hello

121

Table 4.3.

Twenty Most Frequently Uttered Vocalizations and Respective Frequencies

Rank Baseline Ignored Denied

1 26 DUW 31 KS 83 wanna

2 26 NWM 22 wanna 69 Cosmo

3 25 you 18 cosmo 49 KS

4 22 here 17 DUW 40 a

5 22 WBI 16 here 35 bird

6 20 I 11 go 33 DUW

7 20 KS 10 come 31 be

8 16 bird 10 no 31 okay

9 16 okay 9 kiss 29 good

10 16 PH 8 betty 28 here

11 15 cosmo 8 up 26 kiss

12 13 are 6 I 25 NWM

13 13 good 6 okay 23 go

14 13 that 5 NWM 22 up

15 12 a 4 cuddle 20 no

16 11 WW 4 MWH 20 that

17 10 go 3 a 18 come

18 10 up 3 be 17 Betty

19 9 Betty 3 Cos 16 bite

20 9 what 3 feathers 14 Jean

122

Table 4.4

Conversation Excerpts Containing Repetition and Redirection

Excerpt A

Cosmo: Wanna come here?

BJ: No i'm busy

Cosmo: Wanna come here

Wanna come here

BJ: No Cosmo I'm busy

Cosmo: We're gonna go to kiss

BJ: KS KS

Cosmo: Wanna come here?

Excerpt B

Cosmo: Cosmo wanna go up?

BJ: No Cosmo stay in cage okay?

Cosmo: Cosmo go up?

Cosmo be a good bird

Okay go up

Here step up

BJ: How are you

Cosmo: Wanna be a

123

Cosmo wanna come here

Okay cosmo wanna go up

Note. Question marks were added post-hoc to illustrate instances of upward tonal inflection. Original transcriptions contained no punctuation.

124

Table 4.5.

Conversation Excerpts Containing Interruption

Excerpt A

Cosmo: Wanna go up here?

BJ: Cosmo stay in cage okay?

Cosmo: Okay

BJ: Cosmo stay in cage

Cosmo: Cos don't bite okay

BJ: Okay cosmo stay in cage please

Cosmo: Wanna go up here

ID be a good bird

Don't bite okay

BJ: Be a good—

Cosmo: Cosmo wanna go up be a good bird Cosmo

Excerpt B

Cosmo: Wanna be a good bird?

Cosmo go up

Be a good bird?

Okay

Cosmo wanna go up?

125

BJ: Cosmo stay in cage okay?

Cosmo: Cosmo wanna go up?

BJ: No Cosmo

Cosmo stay in cage

Cosmo: Be a good bird?

BJ: Cosmo‘s a good good bird—

Cosmo: Okay

Cosmo go up

Note. Question marks were added post-hoc to illustrate instances of upward tonal inflection. Original transcriptions contained no punctuation.

126

Table 4.6.

Conversation Excerpt Containing a Denied then Met Request

Cosmo: Cosmo go up

BJ: No Cosmo stay in cage okay

Cosmo: Wanna be a bird?

BJ: Cosmo be a good bird

Cosmo go

Cosmo: Wanna

Okay

BJ: Okay be a good bird

Be a good bird

Wanna come here?

Here step up

Now be a good bird stay up here okay?

Please thank you

Cosmo: Wanna kiss?

BJ: Stay up Cosmo

Cosmo: Wanna come kiss?

BJ: KS KS KS

I love you

Cosmo: Come here

BJ: Here I am

127

Cosmo: Please

KS KS

BJ: Okay

Okay Betty Jean wanna kiss

KS KS KS

Note. Question marks were added post-hoc to illustrate instances of upward tonal inflection. Original transcriptions contained no punctuation.

128

Table 4.7.

Conversation Excerpts Containing Persistence and Refocusing

Excerpt A

Cosmo: Cosmo wanna go up?

BJ: No Cosmo stay in cage okay?

Cosmo wanna play telephone?

Telephone

Cosmo: You have reached

BJ: You have reached Cosmo

Cosmo: LA

BJ: LA

Cosmo: OOO

Wanna be a good bird

BJ: Yes Cos

Cosmo: ID

BJ: Be a good bird

Cosmo: Wanna be a bird

BJ: Wanna be a bird

Cosmo: Cosmo go up

Excerpt B

Cosmo: Wanna peanut okay

Be a good bird okay go up

129

Wanna be a good bird

BJ: Yes be a good bird stay in cage okay

Cosmo: Wanna peanut okay

Note. Question marks were added post-hoc to illustrate instances of upward tonal inflection. Original transcriptions contained no punctuation.

130

Table 4.8.

Conversation Excerpt Containing “Be a Good Bird” and “Cosmo Don’t Bite Okay”

Cosmo: Hi

BJ: Hi

Cosmo: MWH

BJ: Fine thank you

How are you?

Cosmo: Wanna be a good bird?

Cosmo go up

Be a good bird?

Okay

Cosmo wanna go up?

BJ: Cosmo stay in cage okay?

Cosmo: Cosmo wanna go up?

BJ: No Cosmo

Cosmo stay in cage

Cosmo: Be a good bird?

BJ: Cosmo‘s a good good bird—

Cosmo: Okay

Cosmo go up

Cosmo be a good bird?

Cosmo don‘t bite okay?

131

BJ: Okay

Cosmo: Cosmo wanna go up?

Cosmo go up

Cosmo wanna go up?

Okay

Come here

Here you are

I‘m here

132

Figure 4.1. Illustration of the testing space, including relevant dimensions and furniture.

133

Figure 4.2. Nonrandom (p =.00024) distributions of speech and nonword vocalizations during non-requesting conversation (baseline), immediately following ignored requests, and immediately following denied requests.

134

Figure 4.3. Nonrandom (p =.00021) distributions of social and nonsocial vocalizations during non-requesting conversation (baseline), immediately following ignored requests, and immediately following denied requests.

135

CHAPTER 5

PATTERNS IN PLACEMENT OF ―JOKES‖ WITHIN CONVERSATIONS BETWEEN

AN AFRICAN GREY PARROT AND HER HUMAN CAREGIVER1

______

1 Colbert-White, E. N., to be submitted

136

Abstract

Humor can strengthen relationships, redirect conflict, and repair a relationship during conflict. Manipulation of a social partner‘s affect through the use of humor has been studied minimally in young children, and not at all in nonhumans (presumably due to similar methodological issues). The current study assessed whether one nonhuman speaker, an African Grey parrot named Cosmo, used vocalizations classified as ―jokes‖ more often during times of relationship conflict. Relationship conflict was defined as times when Cosmo‘s owner, Betty Jean, denied a request made by Cosmo. Two corpora

(i.e., bodies of text) were created from conversations between Cosmo and Betty Jean.

The first corpus contained Cosmo‘s vocalizations immediately preceding requests. The second corpus contained Cosmo‘s vocalizations following refusals. When one outlier was removed, results showed that the distribution of jokes in the pre-request and post- refusal corpora was equal. However, Cosmo‘s overall vocabulary in the two corpora was different, suggesting she did differentiate between the conditions. The results support previous findings with Cosmo that her vocalizations are heavily dependent upon Betty

Jean‘s, and introduce the possibility that the words in Cosmo‘s repertoire serve both literal (e.g., ―Wanna peanut‖ means ―I want a peanut‖) as well as social (e.g., ―Wanna peanut‖ means ―Pay attention to me‖) functions.

Keywords: Humor, African Grey parrot, Psittacus erithacus, laughter, jokes, relationship conflict, affect manipulation

137

Introduction

Instances of language use during information transfer from sender to receiver are termed speech acts. All speech acts have three components: a locutionary component, which is the literal meaning of the utterance (e.g., ―It is cold in here‖ means ―The room temperature is cold‖); an illocutionary component, which is the speaker‘s intention behind the utterance (e.g., ―It is cold in here‖ really means ―Close the window‖); and a perlocutionary component, which is the effect the speaker‘s utterance has on the affect or behavior of the receiver (Austin, 1962). Speech acts can be studied at the level of structure by approaching language from a syntactic or grammatical stance, or from a more functional stance of what the communicative interaction affords the sender (e.g.,

Owings and Morton, 1998; Premack, 1986; de Saussure, 1916/1983; Tomasello, 2003).

Along with manipulative devices such as coercion, bargaining, and persuasion, humor is another way that speech acts can be used for a function other than their literal

(i.e., locutionary) meaning (Vickers, 1974). Humor is considered a social phenomenon, used to strengthen good relationships and to repair damaged ones (Fine & de Soucey,

1977). This may account for why laughter—a social vocalization—occurs infrequently when individuals are alone (Provine & Fischer, 1989). According to Fine & de Soucey

(1977), the act of telling a joke has a structural component (e.g., pun, irony) which facilitates a social function (e.g., promoting inclusion and/or strengthening bonds). This social function of humor as a repairing mechanism during times of relationship conflict was the subject of the current investigation.

During times of conflict, the introduction of humor can have either positive or negative effects at a variety of levels. When a joke is received well, humor can end

138

conflict; when received poorly, a joke can redirect tension away from the original conflict and onto the ill-received joke or pun (Goffman, 1976; 1981). Humor can be relevant to the conflict‘s topic or it can be used strategically to change the topic (Norrick & Spitz,

2008). Humor can be used to respond to criticism to save face (Forester, 2004) or to reframe issues during negotiation (Maemura & Horita, 2012; Sclavi, 2008). Further,

Schwitalla (1987, in Norrick & Spitz, 2008) described the mitigating effect that well- received humor can have on the intensity with which utterances are vocalized between social partners. In all, humor can transform a tense social interaction into a lighter, more manageable one.

In the literature, many studies have observed and reported on humor production, appreciation, and responsiveness in young children (e.g., Groch, 1974; McGhee, 1971;

Zigler, Levine, & Gould, 1967). For example, Groch (1974) observed nursery school children and reported sex differences in their use of and response to humor, as well as the types of humor (e.g., hostile vs. non-hostile) that children find funny. However, there are few empirical or observational studies on children spontaneously using humor to manipulate the affect or behavior of others during times of relationship conflict. In one study with male, Dutch-Moroccan juvenile offenders between the ages of 14-18 years, van Nijnatten and Stevens (2011) reported the use of laughter by one boy following a negative comment to his probation officer. The authors concluded that the function of the laughter was to prevent conflict and to maintain a positive social environment.

Norrick and Spitz (2008) described spontaneous interactions between 3.5-year old children and their use of humor to diffuse aggression. In transcriptions of siblings and friends at play, Norrick and Spitz provided evidence for humor and laughter alone ending

139

conflicts between young children. However, since it is impossible to ask a very young child to self-report on motivations behind using humor, understanding its use as a manipulation tool is difficult. This testability problem may account for why there is no literature describing the earliest age at which young children begin using humor intentionally to resolve conflicts.

Among nonhuman species, play (e.g., Burghardt, 2005), grooming (e.g., Henzi &

Barrett, 1999), and antiphonal duetting (e.g., Seibert, 2006) are all examples of behaviors which serves the function of strengthening social relationships. However, examples of nonhumans understanding or engaging in humor have not been documented. Likewise, as with identifying functional uses of humor in children, empirically observing humor in nonhuman species is untestable. Equally untestable is whether language-trained nonhumans such as Kanzi the bonobo (Pan paniscus) or Alex the African Grey parrot

(Psittacus erithacus) have a concept of using jokes or humor to manipulate social partners.

Difficulty arises in attaching intention or motivation to the behavior of an individual that cannot self-report. Thus, in order to assess objectively a nonhuman speaker‘s use of vocalizations which, in English, are considered jokes, the current study approached humor from an associative perspective. A joke was considered any vocalization that served the function of changing the affect of another, regardless of the sender‘s intention or motivation behind the vocalization. One female African Grey parrot was recorded during conversations with her caregiver. Her vocal output during normal social interaction was compared to her vocal output following times of relationship conflict. In human literature, ―conflict‖ in a relationship has been defined as a series of at

140

least three back and forth turns in which social partners challenge one another (e.g.,

Coulter, 1990; Norrick & Spitz, 2008). For the current study, relationship conflict was operationally defined as the parrot, Cosmo, making a request (e.g., to get out of her cage) and having that request be refused by her caregiver, Betty Jean.

Parrot social behavior includes a strong emphasis on monitoring and strengthening social relationships with conspecifics (e.g., Nottebohm, 1972; Seibert,

2006; Serpell, 1981). Further, previous research with Cosmo has shown that she very carefully monitors Betty Jean‘s location and behavior and modifies her vocalizations appropriately (e.g., Colbert-White, Covington, & Fragaszy, 2011, see also Chapters 3 and

4 of this dissertation). Given Cosmo‘s differential response to Betty Jean‘s location and behavior, I predicted that Cosmo might also closely monitor Betty Jean‘s affect. If this were the case, she might have learned to strategically utter vocalizations previously associated with laughter or praise from Betty Jean (e.g., ―Telephone for bird‖) to manipulate BJ during times of relationship conflict. Assuming Cosmo viewed having a request refused as a relationship conflict, I hypothesized that the distributions of jokes in

Cosmo‘s vocalizations to BJ immediately prior to requests and immediately following denied requests would be nonrandom.

Method

Subject and Housing

Cosmo was 6 years old at the beginning of the project in 2007. Cosmo‘s female caregiver Betty Jean (hereafter BJ) purchased her from a pet store in 2002 when she was

5 months old. Although Cosmo had some experience hearing other human speakers, BJ was her consistent companion. BJ established a simplified grammar with limited

141

vocabulary by labeling new objects for Cosmo and correcting Cosmo‘s misuse and mispronunciation of words. Social interaction with BJ was the primary means by which

Cosmo acquired melodies, English speech, and some nonword sounds like kiss noises.

Additional nonword sounds such as microwave beeps were acquired naturally. Two female dogs also resided with BJ and Cosmo at the time of data collection.

All video-taping was conducted at BJ‘s home with Cosmo in her primary cage

(55.9 x 61.0 x 83.8 cm, with perch extending 40.6 cm from top) which was located in a sun room facing BJ‘s reading chair 2 m away. Food and water were provided ad libitum throughout all testing.

Recording and Transcriptions

For three weeks prior to testing, Cosmo was habituated to a Sony DCR-TRV39 mini-DV video camera on a tripod 1.5 m from the cage. Cosmo‘s cage was the only object in the camera frame. The camera‘s built-in microphone (32 kHz, 16-bit audio) recorded all audio. After the habituation phase, BJ recorded five, hour-long events during two social conditions when Cosmo seemed particularly talkative. During the taping, BJ sat in her reading chair and interacted with Cosmo as normal in the room with her. An illustration of the layout of the space is shown in Figure 5.1.

Videos were transcribed using the code ―ID‖ (i.e., indistinguishable) whenever vocalizations were not clear. Syllables and fragments were transcribed as they were heard (e.g., ―tele‖ and ―showe‖ as in ―telephone‖ and ―shower,‖ respectively). To decrease experimenter bias, contextual information in the videos was not used to construct the transcriptions. Nonword sounds were transcribed with two- or three-letter codes (refer to Table 5.1). Footage used for analysis contained 180 minutes of

142

transcriptions in each condition. An independent observer transcribed approximately 5% of the tapes. ―Matching‖ required both timestamps and vocalizations to be identical between the transcribers. The Cohen‘s kappa coefficients of reliability for the seven reliability samples ranged from κ = .38-.79. When the .38 outlier was removed, the median kappa coefficient rose from .67 to .72.

The SAE Phrase Frequency Tool (Strategic Analysis Enterprises, Inc.,

Williamsburg, VA) computer program searched for recurrent phrases, and not only single words. This program is used by linguists to find recurrent phrases in text. The program tabulated the number of occurrences of all words and phrases ranging one and nine words long (nonword sounds were considered words) that occurred at least once. To aid with qualitative analysis of the conversations, utterances that only occurred once were included, despite the inability to make comparisons across conditions. If a phrase only occurred as part of a larger phrase (e.g., ―in a car‖ was only uttered as ―go in a car‖), then the program tabulated the larger of the two phrases. Phrases were not included if they contained or spanned periods.

Procedure

To create a list of Cosmo‘s ―jokes,‖ a keyword search in AntConc (version

3.2.1w, Laurence Anthony, Waseda University) was conducted to locate all instances of

BJ laughing (tagged ―LA‖ in transcripts). AntConc is freeware program used by linguists to analyze corpora. A corpus is uploaded to the program where operations such as concordances, keywords in context searches, and word counts can be performed.

Cosmo‘s vocalization immediately prior to BJ‘s laughter was recorded. This ―raw‖ jokes

143

list was given to BJ with the instruction to identify all vocalizations that she considered to be jokes. These vocalizations comprised the final joke list (see Table 5.2).

Transcriptions were prepared for analysis using AntConc. Keyword searches identified all of Cosmo‘s denied requests. The 2 min before Cosmo‘s request and the 2 min following BJ‘s refusal were extracted, and two corpora (i.e., bodies of text) were created. Because Cosmo frequently made additional requests following an initial request

(whether it was refused or not), and because BJ often continued to refuse Cosmo‘s requests after the initial refusal, the 2 min often comprised additional requests and refusals. However, the 2 min segments were divided according to the request which was first refused by BJ. An example of this is shown in Table 5.3. Further, to reduce the possibility of replication of transcription content, there was no overlap of excerpts. That is to say, if the 2 min before a targeted refusal event excerpt fell within the 2 min after refusal of a previous refusal event, the second refusal was not included in the analysis.

Data Analysis

A Spearman‘s rho correlation was calculated to determine whether the 20 most commonly uttered vocalizations were correlated (i.e., Did Cosmo use the same vocalizations with similar frequencies before and after refusal events?). A chi-square goodness of fit test was conducted to compare the pre-request corpus to the post-refusal corpus. If Cosmo used jokes randomly, then the expected proportion of jokes in the pre- request and post-refusal corpora would be equal (i.e., .5). All analyses were evaluated at the two-tailed, α = .05 level.

144

Results

Preliminary Observations

Within the five sessions of video footage, 14 refused request events fit the operational definition of no overlap. The 14 refused requests contained five different types of requests: Go up (i.e., get out of cage, requested 3 times), Come here (i.e., approach cage, requested 7 times), Wanna cuddle (requested once), Wanna peanut

(requested twice), and Wanna shower (requested once). These data are organized in

Table 5.4.

The corpus containing 2-min segments before requests was comprised of 246 units (i.e., stand-alone words, phrases, or nonword sounds) by Cosmo. The corpus of 2- min segments following refusals contained 259 utterances. For comparison, BJ uttered

115 and 104 units in the before requests and after refusal corpora, respectively. All analyses were conducted with BJ‘s utterances removed. Cosmo‘s most frequently uttered vocalizations in the before request corpus were KS [kiss sound] (44 times), ―wanna‖ (44 times), ―Cosmo‖ (31 times), KS KS (19 times), and ―bird‖ (18 times). ―ID‖ (i.e., indistinguishable vocalization) and ―a‖ did appear in the top five, but were not included in this list. In the after refusal corpus, the five most commonly uttered vocalizations were

―wanna‖ (46 times), ―Cosmo‖ (38 times), ―go‖ (33 times), KS (26 times), and ―up‖ (26 times). ―ID‖ also appeared in the top five, but was removed.

The 20 most frequently uttered vocalizations for both corpora appear in Table 5.5.

The Spearman‘s rho correlation revealed that the rankings of the most commonly uttered vocalizations were not correlated, ρ(18) = 0.29, p = .19. This indicated that Cosmo‘s overall vocabulary was different before requests and after refusals.

145

Jokes comprised very small percentages of Cosmo‘s vocalizations both before requests (4.1%) and after refusals (5.4%). Of the 11 jokes comprising the joke list, only 8 were represented in the dataset. ―Cosmo‘s a doggie,‖ ―No‖ mimicked in BJ‘s voice, and

―Cosmo don‘t be good Cosmo‖ were not uttered during any of the requesting events. The most frequently uttered joke was ―Cos let go,‖ which was uttered 10 times, followed by a mimic of BJ‘s laugh, which was uttered 7 times (see Table 5.6). As shown in Table 5.6,

―Telephone for bird‖ appeared in the most taped sessions (three out of five), making it the most commonly used joke. Cosmo made 14 jokes following being denied a shower, which was the greatest number of post-refusal jokes. ―Come here‖ had 5 jokes following refusal, ―Go up‖ had 4, ―Wanna peanut‖ had 1, and ―Wanna cuddle‖ had 0.

Distribution of Jokes Pre- and Post-Refusal

Between the two corpora, 34 total jokes were uttered. Ten jokes were uttered in the before request corpus; 24 jokes were uttered in the after refusal corpus. A chi-square goodness of fit test was conducted, with 17 jokes being the expected frequency if the distribution of jokes was equal. The chi-square test determined that the distribution of jokes was not equal, X2(1) = 5.76, p = .016. While the majority of the 14 requesting events contained 0-3 joke utterances, during one requesting event, Cosmo uttered 16 jokes. Fourteen of those jokes were after she was refused, and were part of a repeated dialogue with BJ (see Table 5.7). When this outlier event was removed, the summed frequencies for the before request and after refusal corpora were reduced to 8 and 10 jokes, respectively (see Figure 5.2). To confirm that this new observed distribution was not different from chance, a second chi-square goodness of fit test was calculated, X2(1) =

0.22, p = .64. The data from this second, nonsignificant analysis appeared to be more

146

representative of Cosmo‘s typical joking behavior, which meant that Cosmo‘s use of jokes before and after relationship conflict was equal.

Discussion

The current study explored one African Grey parrot‘s use of vocalizations which, for humans, would be called jokes during times of relationship conflict (i.e., request refusal). For humans, humor functions to strengthen relationships and to repair relationships during times of tension (Fine & de Soucey, 1977). Since there is no empirical evidence of nonhuman species having a concept of humor, and since determining the intention behind Cosmo‘s speech vocalizations is impossible, humor was simplified and approached from an associative framework. In this way, jokes were considered any utterance which, when vocalized, positively influenced BJ‘s affect. When presented with a relationship conflict (i.e., denied request), manipulating BJ‘s affect could (1) decrease the tension associated with a relationship conflict, and/or (2) increase the probability that BJ acquiesces to Cosmo‘s request. Given parrots‘ pre-established use of vocalizations to promote and strengthen bonds among individuals (e.g., Nottebohm,

1972; Seibert, 2006), I predicted Cosmo would be attentive to the introduction of a relationship conflict and strategically use jokes for one of the two above functions.

Contrary to what was hypothesized, the analysis revealed that Cosmo uttered jokes before making requests just as often as she uttered jokes after her requests were denied.

For Cosmo, despite repeated pairings, BJ‘s laughter does not appear to be associated with a positive affect the way that it is in humans (e.g., Hayworth, 1928).

From the associative framework underlying this study, the results could be explained as

Cosmo having experienced too few pairings of BJ‘s laughter with positive behaviors

147

(such as giving treats or meeting requests) to learn the value attached to laughter. Thus, while previous work (e.g., Colbert-White et al., 2011) and Chapters 3 and 4 of this dissertation confirm that Cosmo does closely monitor and respond to BJ‘s behavior, and though according to BJ, Cosmo makes jokes to strengthen their relationship (personal communication, 2010), BJ‘s laughter does not appear to be a reliable, informative stimulus. In humans, the use of humor and laughter to diffuse relationship tension is argued to develop before the age of 4 (e.g., Norrick & Spitz, 2008). Thus, children as young as 3 years old are able to pick up on socioemotional cues which seem imperceptible to Cosmo.

The study‘s null findings may also—or instead—indicate that Cosmo does not view refused requests as relationship stressors. Instead, more salient events such as scolding (e.g., after Cosmo is caught ripping BJ‘s wallpaper) may be perceived by Cosmo as disruptions to their social relationship. To assess this hypothesis, additional recordings during scolding events would be necessary. Likewise, the 11 utterances defined as jokes because they made BJ laugh may not serve any special function for Cosmo beyond continuing social interaction with BJ. For Cosmo, joke utterances may be a reliable attention-drawing stimulus, as they are likely to solicit a response (i.e., laughter) from BJ.

This could explain why they are just as likely to be uttered before a request as after a refusal. In both situations jokes would maintain or solicit conversation; serving to keep

BJ‘s attention in the minutes leading up to a request, as well as in the minutes following a refusal. Following refusals, jokes might maintain BJ‘s attention so that the request could be reframed or repeated—which was demonstrated in Chapter 4.

148

Cosmo‘s use of vocalizations such as ―Cosmo wanna poop‖—or any vocalization in her repertoire, for that matter—may not always be uttered for literal purposes. Instead, they may serve the function of prolonging a conversation. Such a finding would still indicate Cosmo uses speech to manipulate BJ, but it would mean manipulation of BJ‘s time rather than her affect. Even if Cosmo does not perceive refused requests as relationship conflicts, she does appear to strategically organize her utterances so as to get

BJ‘s attention and maintain it differentially depending upon whether she is preparing to request something of BJ or after she has been refused. This shows that Cosmo does monitor closely what BJ says in response to own vocalizations, and she has a concept of what it means to have a request refused.

The current study‘s findings are supported not only by the results of Chapter 4, but also by previous work with other social nonhumans. Chimpanzees (Pan troglodytes,

Call, Hare, Carpenter, & Tomasello, 2004), capuchin monkeys (Cebus apella, Phillips,

Barnes, Mahajan, Yamaguchi, & Santos, 2009), and even African Grey parrots (Péron,

Chardard, Nagle, & Bovet, 2011) respond differentially to a human partner being unwilling or unable to provide a desired object. Such ability involves monitoring the partner‘s physical behavior and responding to nonverbal behavioral cues provided by the partner. Previous research with Cosmo has determined that at least one African Grey can use verbal cues to make judgments about a social partner‘s willingness or unwillingness to engage in vocal interaction (Colbert-White et al., 2011). While Greys demonstrate the ability to respond to human social partners‘ nonverbal, physical cues as well as verbal cues, human sociocognitive cues (i.e., laughter, positive tone) may be too species-specific

149

to humans for Cosmo to learn to attend to, even after 7 years of having BJ as her primary social partner.

In the future, I would like to tease apart the possible roles that interspecific sociocognitive abilities and stimulus–response pairing play in Cosmo‘s lack of use of vocalizations associated with BJ‘s laughter and positive affect. In humans, intonation

(i.e., tone of voice), like laughter, can indicate valuable information about an individual‘s affect (e.g., Stern, Spieker, & MacKain, 1982). Assessing African Greys‘ response to rising (i.e., positive) and falling (i.e., negative) intonation in a forced-choice task would address one type of vocal, but nonverbal human social cue. I would also like to follow up on the current study by returning to the dataset and exploring the differential use of specific vocalizations. For example, Cosmo used kiss sounds numerically more often after she was refused compared to before she made requests. While there was no overall difference in joke usage, there may be other vocalizations that Cosmo uses with differing frequencies following particular conversational events like refusals. Rather than using

BJ‘s laughter to establish a list of utterances to explore, I would like to look for vocalizations immediately prior to met requests (e.g., when BJ does let Cosmo out of her cage), and examine the use of these particular vocalizations‘ within conversations. Such a list of vocalizations may provide additional evidence for strategic placement of vocalizations by Cosmo to manipulate BJ‘s behavior.

Being able to monitor and respond to changes in a social partner‘s behavior and affect are skills that contribute to strong relationships. While the current study‘s findings did not demonstrate increased use of jokes following a presumed relationship conflict, they did support another study‘s (Chapter 4) finding that one African Grey finely-tuned

150

her vocalizations in response to having a request refused. Specifically, the parrot differentially used vocalizations in her repertoire to prolong conversation with her owner both during the time leading up to requests and following refusals. This functional use of vocalizations beyond their literal English meaning implies that the bird‘s repertoire may simultaneously serve two functions: literal and social.

151

Acknowledgements

Hannah Hall and Marisol Macias both aided with literature searching. In addition, I thank Hannah Hall and Allison Kaufman for assisting with reliability coding and associated calculations, respectively. I also thank Dominic Byrd for his illustration of the testing area in Figure 5.1.

152

References

Austin, J. L. (1962). How to do things with words. Cambridge, MA: Harvard University

Press.

Burghardt, G. (2005). The genesis of animal play: Testing the limits. Cambridge, MA:

MIT Press.

Call, J., Hare, B., Carpenter, M., & Tomasello, M. (2004). ‗Unwilling‘ versus ‗unable‘:

Chimpanzees‘ understanding of human intentional action. Developmental

Science, 7, 488-498. doi:10.1111/j.1467-7687.2004.00368.x

Colbert-White, E. N., Covington, M. A., & Fragaszy, D. M. (2011). Social context

influences the vocalizations of a home-raised African Grey parrot (Psittacus

erithacus erithacus). Journal of Comparative Psychology, 125, 175-184.

doi:10.1037/a0022097

Coulter, J. (1990). Elementary properties of argument sequences. In G. Psathas (Ed.),

Interaction competence (pp. 181-204). Washington DC: University Press of

America. de Saussure, F. (1983). Course in general linguistics. Chicago: Open Court

Publishing. (Original work published 1916)

Fine, G. A, & de Soucey, M. (1977). Joking culture: Humor themes as social regulation

in group life. Humor, 18, 1-22. doi: 0933-1719/05/0018-0001

Forester, J. (2004). Responding to critical moments with humor, recognition, and hope.

Negotiation Journal, 20, 221-237. doi:10.1111/j.1571-9979.2004.00019.x

Fouts, R. (1997). Next of kin: My conversations with chimpanzees. New York:

HarperCollins Publishers, Inc.

153

Goffman, E. (1976). Replies and responses. Language in Society, 5, 257-313.

Goffman, E. (1981). Forms of talk. Philadelphia: University of Pennsylvania Press.

Groch, A. S. (1974). Joking and appreciation of humor in nursery school children. Child

Development, 45(4), 1098-1102.

Hayworth, D. (1928). The social origin and function of laughter. Psychological Review,

35, 367-384. doi:10.1037/h0073133

Henzi, S. P., & Barrett, L. (1999). The value of grooming to female primates. Primates,

40, 47-59. doi:10.1007/BF02557701

Maemura, Y., & Horita, M. (2012). Humour in negotiations: A pragmatic analysis of

humour in simulated negotiations. Group Decision and Negotiation, 21, 821-838.

doi:10.1007/s10726-011-9251-9

Norrick, N. R., & Spitz, A. (2008). Humor as a resource for mitigating conflict in

interaction. Journal of Pragmatics, 40, 1661-1686.

doi:10.1016/j.pragma.2007.12.001

Nottebohm, F. (1972). The origins of vocal learning. The American Naturalist, 106, 116–

140.

Owings, D., & Morton, E. (1998). Animal vocal communication: A new approach.

Cambridge: Cambridge University Press.

Péron, F., Rat-Fischer, L., Nagle, L., & Bovet, D. (2010). ‗Unwilling‘ versus ‗unable‘:

Do grey parrots understand human intentionally action? Interaction Studies, 11,

428-441. doi:10.1075/is.11.3.06per

154

Phillips, W., Barnes, J. L., Mahajan, N., Yamaguchi, M., & Santos, L. R. (2009).

‗Unwilling‘ versus ‗unable‘: Capuchin monkeys‘ (Cebus apella) understanding of

human intentional action. Developmental Science, 12, 938-945.

doi:10.1111/j.1467-7687.2009.00840.x

Premack, D. (1986). “Gavagai” or the future history of the animal language controversy.

Cambridge, MA: MIT Press.

Provine, R. R., & Fischer, K. R. (1989). Laughing, smiling, and talking: Relation to

sleeping and social context in humans. Ethology, 83, 295-305.

doi:10.1111/j.1439-0310.1989.tb00536.x

Sclavi, M. (2008). The role of play and humor in creative conflict management.

Negotiation Journal, 24, 157-180. doi:10.1111/j.1571-9979.2008.00175.x

Seibert, L. M. (2006). Social behavior of Psittacine birds. In A. U. Luescher (Ed.),

Manual of parrot behavior (pp. 43-48). Ames: Blackwell Publishing.

Serpell, J. (1981). Duets, greetings and triumph ceremonies: Analogous displays in the

parrot genus Trichoglossus. Zeitschrift Tierpsychologie, 55, 268–283.

doi:10.1111/j.1439-0310.1981.tb01272.x

Stern, D. N., Spieker, S., & MacKain, K. (1982). Intonation contours as signals in

material speech to prelinguistic infants. Developmental Psychology, 18, 727-735.

doi:10.1037/0012-1649.18.5.727

Tomasello, M. (2003). Constructing a language: A usage-based theory of language

acquisition. Cambridge, MA: Harvard University Press.

155

van Nijnatten, C., & Stevens, G. (2011). Juvenile participation in conversations with

probation officers. International Journal of Offender Therapy and Comparative

Criminology, 56, 483-499. doi:10.1177/0306624X11399872

Vickers, C. H. (1974). Levels of human communication. Communication, 1(1), 7-24.

Zigler, E., Levine, J., & Gould, L. (1967). Cognitive challenge as a factor in children‘s

humor appreciation. Journal of Personality and Social Psychology, 6, 332-336.

doi: 10.1037/h002429

156

Table 5.1.

Nonword Sound Coding Scheme

CODE DESCRIPTION CODE DESCRIPTION

AM Telephone answering machine beep LS Laser sound

CR Crow caw LSW ―Laser sound-ID-whistle‖ combination

DS Any dog bark, gruff, or howl MWH Miscellaneous one-note whistle

DSS Any dog bark, gruff, or howl sequence NWM Miscellaneous non-whistle

DO Door opening creak NWMS Miscellaneous non-whistle sequence

DOS Door opening creak sequence OOO ―Ooh‖ and other long-o sounds

DUW Duet whistle OU ―Oww‖ as in pain

DUWS Duet whistle sequence OW Owl hoot

DW Dog whine/whimper PH Telephone dialing beep

DWS Dog whine/whimper sequence PHS More than five telephone dialing beeps

FR Frog croak RI Telephone ring

157

HA Hawk vocalization RIS Telephone ring sequence

HAS Hawk vocalization sequence WBI Wild miscellaneous songbird

ID Indistinguishable WBIS Wild miscellaneous songbird sequence

―Woo-woo-woo‖ or ―woo‖ (B. J.‘s KS Kiss sound WF vocalization of a dog barking)

KSS Kiss sound sequence WW Wolf whistle

LA Laugh WWS Wolf whistle sequence

Note: ―Sequence‖ = Vocalization repeatedly uttered for more than 4 seconds. ―Duet whistle‖ = Melodic whistling, not natural bird vocalization.

158

Table 5.2.

List of Cosmo’s “Jokes” Used in Analyses

Telephone for bird

You have reached Cosmo

Cosmo wanna poop

Cosmo‘s a doggie

Mary has feathers

Goodbye

Cosmo wanna go for a walk

Cosmo don‘t be good Cosmo

[Mimic of BJ‘s laugh]

Cos let go

Note. ―Jokes‖ were determined in a two-step process. ECW recorded Cosmo‘s vocalizations occurring immediately prior to BJ‘s laughter. This list was given to BJ, who made the final selection of which vocalizations she considered Cosmo‘s ―jokes.‖ For clarification purposes, Mary is one of BJ‘s dogs; and both the ―Goodbye‖ and ―No‖ vocalizations are uttered in BJ‘s voice during a time that, for humans, would be considered ironic or sarcastic (e.g., BJ: ―Cosmo wanna play telephone?‖ Cosmo:

―Goodbye‖).

159

Table 5.3.

Example Data Preparation for Multiple Requests

Cosmo: Wanna peanut?

BJ: No

No more peanut

Cosmo: Cosmo wanna be a good bird?

BJ: Yes be good bird

Cosmo: Cosmo go up here

Cosmo go up

Go up ID okay

BJ: Cosmo stay in cage okay?

Note. Cosmo‘s request ―Wanna peanut‖ was the target request for which the 2 min prior and 2 min following BJ‘s ―No‖ were compared. Despite Cosmo‘s three requests to ―go up‖ (i.e., get out of her cage), and BJ‘s refusal, these utterances were included in the 2 min following the ―Wanna peanut‖ target request. Question marks were added post-hoc to illustrate instances of upward tonal inflection. Original transcriptions contained no punctuation.

160

Table 5.4.

Distribution of Jokes Across Conflict Events for Each Taped Session

JOKES

Sess. Request Before Request After Refusal

1 Go up Telephone for bird Telephone for bird

Go up You have reached Cosmo (2x) LA

Cosmo wanna poop

Go up N/A LA

Come here Mary has feathers (3x) N/A

2 Wanna cuddle N/A N/A

Come here Mary has feathers N/A

Come here Mary has feathers Telephone for bird

3 Wanna peanut N/A N/A

4 Wanna shower Goodbye LA (4x)

LA Cos let go (10x)

5 Wanna peanut N/A Cosmo wanna go for a walk

Come here N/A N/A

Come here N/A Telephone for bird (2x)

Come here N/A Cosmo wanna go for a walk

Note. ―Sess.‖ refers to each taped session. Numbers in parentheses represent repeated use of a particular joke. ―N/A‖ indicates no jokes were uttered a particular 2-min segment.

161

―LA‖ = Mimicry of BJ‘s laugh. ―Goodbye‖ is considered a joke by BJ when it is mimicked in her voice.

162

Table 5.5.

Twenty Most Frequently Uttered Vocalizations and Associated Frequencies

Rank Before Request After Refusal

1 44 KS 46 wanna

2 44 wanna 38 cosmo

3 31 cosmo 33 go

4 19 KS KS 26 KS

5 18 bird 26 up

6 17 DUW 23 bird

7 17 go 20 DUW

8 17 good 17 go up

9 15 NWM 17 good

10 15 up 16 NWM

11 14 no 16 okay

12 14 okay 15 cosmo wanna

13 13 betty 15 here

14 13 MWH 14 bird

15 12 here 12 kiss

16 12 kiss 11 KS KS

17 11 feathers 11 let

18 11 go up 11 no

19 10 be a good bird 10 be a good bird

20 10 cosmo wanna 10 let go

163

Note. Indistinguishable codes (―ID‖) and articles (i.e., ―a,‖ ―an,‖ ―the‖) were removed.

164

Table 5.6.

Joke Use Frequencies

Joke Frequency # of Sessions

Telephone for bird 5 3

You have reached Cosmo 2 1

Cosmo wanna poop 1 1

Cosmo‘s a doggie 0 N/A

Mary has feathers 5 2

Goodbye 1 1

Cosmo wanna go for a walk 3 1

No 0 N/A

Cosmo don‘t be good Cosmo 0 N/A

[Mimic of BJ‘s laugh] 7 2

Cos let go 10 1

Note. The mimic of BJ‘s laugh appears in the text as ―LA.‖ There were five recording sessions.

165

Table 5.7.

Outlier Requesting Event from Session 4

Cosmo: Cosmo wanna shower?

BJ: No no shower

Cosmo: Here step up

Step up up up up up up

Let go

OOO what a bird

BJ: LA

Cosmo: Let go

Let go

Let go LA

BJ: LA

Cosmo: let

Let go

Let go please

Let go

BJ: LA

Cosmo: Let go

Let go

166

BJ: LA

Note. The question mark was added post-hoc to illustrate an instance of upward tonal inflection. Original transcriptions contained no punctuation. ―LA‖ = Mimic of BJ‘s laughter.

167

Figure 5.1. Illustration of the testing space, including relevant dimensions and furniture.

168

Figure 5.2. Total number of ―jokes‖ vocalized during requesting events following removal of one outlier. Before Request is 2 min of utterances before a denied request was made; After Refusal is 2 min of utterances following refused requests. Thirteen requesting events were included in the analysis. The distribution of jokes in both corpora was not different from what would be expected by chance (dotted line), p = .64.

169

CHAPTER 6

GENERAL CONCLUSIONS

For social species, forming, maintaining, and strengthening bonds with others is critical. Beginning during infancy with basic, non-linguistic features like joint attention

(Carpenter, Nagell, Tomasello, Butterworth, & Moore, 1998), children learn how to navigate social interactions to regulate relationships with others. This learning typically ends during adolescence as children refine skills such as persuasion (Clark & Delia,

1976). Like humans, parrots are also highly social. Parrots bond strongly with one individual (e.g., Nottebohm, 1972), are highly affiliative with conspecifics (e.g., Seibert,

2006), have a complex fission-fusion social system (e.g., del Hoyo, Elliott, & Sargatal,

1992), and at least one species is said to engage in aerial play and ―social enjoyment‖

(cockatiel, Nymphicus hollandicus, Cramp, 1994). The vocal repertoire of wild parrots includes a variety of calls specifically used for the purposes of promoting social cohesion and group movement (Bradbury, 2003). Similar to humans, parrots learn their vocalizations from parents and flockmates (Nottebohm & Nottebohm, 1969; Nottebohm,

1970; Pepperberg, 1985, 1994). The commonality of vocalization learning and social behavior in parrots and humans, along with my previous work with Cosmo (Colbert-

White, Covington, & Fragaszy, 2011), led me to predict that one African Grey parrot‘s use of vocalizations during conversations with her caregiver might show evidence of learned human features of conversations. In a series of four studies, I investigated turn-

170

taking and thematic cohesion, auditory perceptual perspective taking, requesting behavior, and affect manipulation.

Summary of Findings from Present Studies

Evidence for Thematic-Linked Conversation

In Chapter 2, a qualitative and quantitative review of conversations (defined based on time between turns) between Cosmo and BJ confirmed that Cosmo frequently initiates and maintains thematically-linked conversations with BJ. Bouts of turn-taking ranged from 2-9 turns, all within a predetermined window of 3 s per turn (within or between speakers). Themes centered largely around initiating and prolonging vocal contact, and

Cosmo initiated conversation with BJ much more frequently than BJ initiated with her.

Human-like conversational features were apparent. Though BJ and Cosmo typically honored turn-taking rules, there were instances of both speakers interrupting and talking over each other. In one recording, during a 2-min back-and-forth event where Cosmo wanted to be let out of her cage, Cosmo interrupted BJ (―Cosmo is a good good bir—‖

―Okay / Cosmo go up‖). Given the social context of a continuously denied request,

Cosmo‘s behavior may have been a display of frustration, as is seen with humans

(Schiffrin, 1985).

There was also evidence of Cosmo initiating and maintaining conversations with

BJ which were not relevant only to requests. Terrace‘s (1986) controversial conclusions about his signing chimpanzee Nim Chimpsky were that Nim rarely initiated conversation with his handlers, rarely signed responses other than what his handlers had just signed to him, and did more requesting than conversing. Contrary to this, conversations between

Cosmo and BJ were much more human-like. Cosmo frequently initiated conversation,

171

responded to BJ‘s vocalizations variably depending upon the social or environmental context, and held thematically-linked conversations which did not pertain to requests

(e.g., BJ going to work or having company). While there is strong evidence (Chapters 4 and 5) that many vocalizations in Cosmo‘s repertoire may serve the function of furthering social interaction, the results of this study confirmed that Cosmo can monitor what BJ says and responds appropriately, often for as many as 8 back-and-forth turns. These results strongly oppose those of researchers such as Terrace.

Evidence for Perceptual Perspective Taking

In Chapter 3, I presented results of an auditory perspective taking study. Cosmo vocalized significantly more loudly during times of visual separation from her owner.

That is to say, when the auditory barrier of distance was introduced, Cosmo appeared to understand that she needed to vocalize more loudly in order for her social partner to hear her. Such perspective taking is a skill attributed to human theory of mind (Premack &

Woodruff, 1978). However, not all vocalizations were louder when BJ was in a different room. Many vocalizations were within Cosmo‘s normal range of 65 dB while still others were as low as 50 dB and barely detected by the decibel meter. A secondary analysis evaluated whether the variation in volume was related to whether vocalizations were social or nonsocial in nature. While the analysis found that overall social vocalizations were uttered louder than nonsocial vocalizations, and Cosmo uttered social vocalizations more loudly when her owner was in the room with her, she did not utter social vocalizations more loudly when her owner was out of the room.

I interpreted these inconsistent findings as evidence either that the coding scheme used to classify vocalizations as social and nonsocial was not appropriate, or that all

172

vocalizations can serve a social function as long as BJ is available and willing to respond.

Vocalizations classified as nonsocial such as ―Betty Jean gonna go in a car‖ or even AM

[answering machine beep] could solicit BJ‘s attention, and therefore might be ―as social‖ as vocalizations like ―I love you‖ and ―Cosmo wanna talk‖ which, in English, are clearly social.

Evidence for Communicative Competence

Communicative competence is a measure of individuals‘ sociolinguistic abilities during interactions with social partners. Such abilities are developed over time through frequent social interactions. In Chapter 4, I explored patterns in Cosmo‘s vocal behavior following denied and ignored requests. In the literature, response to unmet requests is age defined. As children age, perspective taking, additional social skills, cognitive skills, and language become mediating factors in the effectiveness of requesting behavior (Stein

& Albro, 2001). For example, by 4 years old, children still rely heavily upon repetition of requests like very young children (Wootton, 1981), but can also provide more justification for requests (Clark & Delia, 1976; Ervin-Tripp, Guo, Lampert, 1990) and suppress frustration after refusals (Ervin-Tripp et al., 1990). These older children also repeat the addressee‘s name before retrying a request that went ignored (Garvey, 1975), and modify their argument strategies in response to the social partner‘s identity (Stein &

Albro, 2001). By age 6, children consider the desires of others, reflecting the emergence of perspective-taking (Axia, 1996; Clark & Delia, 1976). Around this same age, persuasive behavior incorporates compromise and bargaining (Weiss & Sachs, 1991). It is not until adolescence that the greatest attunement of the persuader to others‘ psychological states occurs (Clark & Delia, 1976). Thus, while much of requesting and

173

denied request behavior is learned, predictable patterns do emerge as sociocognitive abilities unfold.

Cosmo‘s behavior when denied mirrored that of very young children. She modified and repeated her requests to BJ, and there may have been some indication of frustration (see also Chapter 5) when Cosmo interrupted BJ to re-request following being denied. Though it is difficult to make a definitive conclusion, Cosmo‘s speech patterns, in English, appear to incorporate persuasion. Cosmo spontaneously vocalized utterances such as ―Cosmo don‘t bite okay‖ and ―Be a good bird‖ during times when her requests were denied, but not when her requests were met. Further, when she was ignored, Cosmo vocalized more social vocalizations (e.g., kiss sounds, ―I love you‖) than when there were no requests or denied requests. While assessing intention behind Cosmo‘s vocalizations is impossible, the results of this study supported Colbert-White et al.‘s (2011) finding that

Cosmo is sensitive to social context and monitors BJ‘s behavior, both physical and vocal.

The data also supported Colbert-White et al.‘s conclusion that Cosmo is aware when she is being ignored, as indicated by differential use of the utterances in her vocal repertoire.

I interpreted these findings as evidence of basic communicative competence. That is,

Cosmo has learned how to modify her vocalizations during conversation with BJ so as to more effectively meet her own goals.

No Evidence for Affect Manipulation

In Chapter 5, I examined Cosmo‘s vocalizations during times of relationship conflict (defined as denied requests by BJ) for evidence of affect manipulation through what, in English, would be considered humor. Humor is a learned device which, at its most basic, can be considered a behavior (physical or vocal) which positively influences

174

the emotional affect of the listener. During times of relationship conflict, the introduction of humor can have positive effects at a variety of levels. Humor can be relevant to the conflict‘s topic or it can be used to strategically change the topic (Norrick & Spitz, 2008).

Humor can be used to respond to criticism to save face (Forester, 2004) or to reframe issues during negotiation (Maemura & Horita, 2012; Sclavi, 2008). Further, Schwitalla

(1987, in Norrick & Spitz, 2008) described the mitigating effect that well-received humor can have on the tone intensity with which utterances are vocalized between social partners. For adults, humor can transform a tense social interaction into a lighter one.

Much less is known about humor use by children. There is some evidence that children as young as 3 use humor to resolve or lessen conflict, but these conclusions must be taken with caution given the difficulty in assessing intention and motivation in young children.

This study was developed based on anecdotal evidence from BJ that Cosmo made

―jokes‖ when she ―got in trouble.‖ While recorded sessions contained no instances of

Cosmo being scolded by BJ, I explored the possibility that having a request refused by BJ might be enough of a relationship conflict to find patterns in joke usage. I searched two corpora, Pre-Request and Post-Refusal, for the 11 vocalizations in Cosmo‘s repertoire that BJ considered jokes. Cosmo was equally likely to use jokes during both speech contexts, but did show overall differences in the selection and placement of vocalizations.

The results of this study did not support the anecdotal evidence for affect manipulation, leading me to believe that refused requests may not signal a relationship conflict for

Cosmo. However, the findings did support previous work with Cosmo (Colbert-White et al., 2011, see also Chapter 4) which demonstrated that Cosmo‘s utterances are strongly dependent upon BJ‘s.

175

The Role of Ecological Relevance

Assessing a nonhuman‘s behavior for evidence of human-like features must be done with caution. For humans, complex conversational strategies such as humor, persuasion, bargaining, and perspective-taking are necessary to communicate effectively while still maintaining social relationships. While these abilities might be helpful for any social species, assumptions about whether a species can learn to use them are speculative at best. While my previous (Colbert-White et al., 2011) and current work allow me to confidently conclude that one African Grey parrot monitors who, what, and where features of her social environment, considering the species‘ natural history is necessary to aid with interpretations of other findings.

In the turn-taking and dialogue study, Cosmo was found to engage in back-and- forth speech and nonspeech duets with BJ. These duets were variable in composition, and occurred more frequently during times when BJ was in the room with her. In the wild, many parrots engage in antiphonal duetting (e.g., Nottebohm, 1972; Serpell, 1981), and some have specific vocalizations for maintaining social contact during times of visual separation. Given duetting‘s strong ecological relevance to parrots, it is no surprise that

Cosmo exhibits this species-typical behavior with BJ. What is interesting, however, is the variation on the vocalizations. Some parrots have specific calls which are unique to a particular pair and serve the function of contact and bonding (e.g., yellow-naped

Amazon, Amazona auropalliata, Wright, 1996). While Cosmo and BJ have such

―signature‖ pair vocalizations where either partner can produce all of the parts (e.g., whistling Bridge Over the River Kwai), both Cosmo and BJ also go back-and-forth with many novel phrases. Even more, within these novel duets, the next speaker‘s utterance is

176

sometimes related thematically to the previous utterance of the partner. Cosmo‘s novel duetting behavior indicates that wild paired parrots may also produce utterances generatively and flexibly in their duets, behaviors which have not been documented.

Many species modify qualities of their vocalizations (e.g., rate, pitch, volume) in response to the environment. In humans, modifying vocalizations to ensure accurate information transfer in the presence of a social partner is termed perspective taking

(Premack & Woodruff, 1978). In perspective taking, individuals realize that what a social partner perceives or knows may be different from what they perceive or know, and respond accordingly. Among those who study nonhuman animal behavior, there are some who entertain the possibility of perspective taking by nonhumans (e.g., Bräuer,

Call, & Tomasello, 2004). Others interpret behavior such as voice quality modification in nonhumans to be indicative of only learned behavioral processes (e.g., Slabbekoorn &

Peet, 2003).

Cosmo‘s spontaneous volume modification during times of visual separation adds to the literature on other nonhumans which suggests that some species can take the perspective of social partners (e.g., chimpanzees, Pan troglodytes, Melis, Call, &

Tomasello, 2006; dogs, Canis familiaris, Call, Bräuer, Kaminski, & Tomasello, 2003;

African Grey parrots, Péron, Rat-Fischer, Nagle, & Bovet, 2010). Such ability would be highly ecologically relevant to any species where knowing what a conspecific can see or hear is particularly valuable. This is especially true for parrots which live and forage in dense forests (del Hoyo, Elliott, & Sargatal, 1992), prefer to maintain physical and vocal contact with pair-mates (Seibert, 2006), and have complex fission-fusion social systems

(Burish, Kueh, & Wang, 2004). For parrots, knowing what others‘ know is a highly

177

beneficial capability. Thus, like with turn-taking, it is no surprise that the results of this study were positive.

The final two studies in my dissertation, response to unmet requests and social partner affect manipulation, are arguably more ecologically relevant to humans than to parrots. For this reason, rather than framing conclusions around human capabilities, I approached these analyses from an associative perspective. Just as young children learn through experience what ―works‖ and what does not ―work‖ when making requests of others, and how to use humor to mitigate tense social interactions, I hypothesized that an

African Grey might also be able to learn these skills. Cosmo did treat being ignored and denied as different situations, and responded accordingly. However, many of the vocalizations in her repertoire appeared to serve either a social-only or social and literal function. That is to say, the sole function of some utterances, regardless of what they mean in English, may only function to solicit attention from BJ or to maintain contact.

Just as wild parrots have a small repertoire (10-15 calls) of vocalizations that serve particular functions (Bradbury, 2003), Cosmo may make little functional distinction among vocalizations which, in English, have very different definitions.

A similar conclusion was drawn in the affect manipulation study. The null findings beg the question of the extent to which a wild parrot pair monitors features beyond physical and vocal behavior. If two humans are conversing, they can do so out of anger, love, or sadness, for example. Emotions during interactions are monitored and very relevant to most humans, which explains why we might try to manipulate them using humor, or perceive them in our companion animals. In nonhumans, observing and measuring affect is very difficult to test, which makes interpreting the results of the jokes

178

study impossible from an emotional stance. However, the data did support previous work

(Colbert-White et al., 2011, Chapters 3 and 4) which showed that Cosmo is highly attuned to BJ‘s behavior.

Future Directions

The current series of studies has illuminated a theme of adaptation. Not only has

Cosmo adapted to having a human as her predominant social partner, she has learned a species atypical communication system. Though there are few data on wild African Grey parrot vocal behavior, I would like to explore further the extent to which the vocal abilities of exceptional parrots like Cosmo mirror wild parrot behavior. The results of my dissertation suggest that wild Greys‘ communication system may include generativity of calls, perceptual perspective taking, and strategic use of discrete vocalizations in a goal- oriented fashion. Field research involving recordings of wild African Greys may provide parallels to what I have found with Cosmo and her species atypical communication system.

In order to replicate and extend the current study‘s perceptual perspective taking findings, one possibility is to repeat the procedure with a closed door, thus providing an alternative auditory barrier to distance. If Cosmo does, in fact, monitor her surroundings and has a concept of what BJ can and cannot hear, additional differences in volume may be observed. Alternatively, artificial noise (e.g., Penna et al., 2005) would present a different type of auditory barrier to which Cosmo would have to adapt in order to be heard by BJ. Given the ecological validity of maintaining contact through barriers in the wild, the results of such a study may provide information about the fine-tuning of perceptual accommodation by African Greys in the wild.

179

Collaboration with computational linguists to develop probabilistic models of

Cosmo‘s conversations with BJ may allow for more objective measures of intentionality.

Without being able to ask Cosmo the intention behind her selection and placement of vocalizations following denied and ignored requests, it is inappropriate to conclude with confidence that any observed patterns are intentional. However, given the 278 vocalizations in Cosmo‘s repertoire (Colbert-White et al., 2011), the probability of particular vocalizations occurring, and then occurring following a particular response from BJ, can be assessed. Developing such a model would strengthen claims about

Cosmo‘s use of her communication system. In particular, intentionality, manipulation, and perhaps behavioral indicators of frustration (e.g., interruption) could be assessed.

Conclusions

To date, most nonhuman language research has focused predominantly on the structural features of language, including syntax and label learning. In a series of four studies, I have demonstrated that the unexplored side of language—functional conversational features—merits further research with nonhumans. Features such as turn- taking, auditory perspective taking, and requesting behavior all require a level of communicative competence. I have shown that species other than humans may demonstrate such abilities if given the opportunity. Further, the results of controlled studies with captive animals like Cosmo may provide valuable information about complex communication systems of wild individuals for which data on species-typical vocalizations may be unavailable.

180

References

Axia, G. (1996). How to persuade mum to buy a toy. First Language, 16, 301-317. doi:

10.1177/014272379601604803

Bradbury, J. W. (2003). Vocal communication in wild parrots. In F. B. M. de Waal, & P.

L. Tyack (Eds.), Animal social complexity: Intelligence, culture, and

individualized societies (pp. 293-316). Cambridge, MA: Harvard University

Press.

Bräuer, J., Call, J., & Tomasello, M. (2004). Visual perspective taking in dogs (Canis

familiaris) in the presence of barriers. Applied Animal Behaviour Science, 88,

299-317. doi:10.1016/j.applanim.2004.03.004

Burish, M. J., Kueh, H. Y., & Wang, S. S.-H. (2004). Brain architecture and social

complexity in modern and ancient birds. Brain, Behavior and Evolution, 63, 107-

124. doi:10.1159/000075674

Call, J., Bräuer, J., Kaminski, J., & Tomasello, M. (2003). Domestic dogs (Canis

familiaris) are sensitive to the attentional states of humans. Journal of

Comparative Psychology, 117, 257-263. doi:10.1037/0735-7036.117.3.257

Carpenter, M., Nagell, K., Tomasello, M., Butterworth, G., & Moore, C. (1998). Social

cognition, joint attention, and communicative competence from 9 to 15 months of

age. Monographs of the Society for Research in Child Development, 63, 1-174.

Clark, R. A., & Delia, J. G. (1976). The development of functional persuasive skills in

childhood and early adolescence. Child Development, 47, 1008-1014.

doi:10.2307/1128437

181

Colbert-White, E. N., Covington, M. A., & Fragaszy, D. M. (2011). Social context

influences the vocalizations of a home-raised African Grey parrot (Psittacus

erithacus erithacus). Journal of Comparative Psychology, 125, 175-184. doi:

10.1037/a0022097

Cramp, S. (1994). Handbook of the birds of Europe, the Middle East, and North Africa:

The birds of the Western Palearctic. Oxford: Oxford University Press. del Hoyo, J., Elliott, J., & Sargatal, J. (1992). Handbook of the birds of the world.

Barcelona: Lynx Edicions.

Ervin-Tripp, S., Guo, J., & Lampert, M. (1990). Politeness and persuasion in children‘s

control acts. Journal of Pragmatics, 14(2), 307-331.

Forester, J. (2004). Responding to critical moments with humor, recognition, and hope.

Negotiation Journal, 20, 221-237. doi:10.1111/j.1571-9979.2004.00019.x

Garvey, C. (1975). Requests and responses in children‘s speech. Jouranl of Child

Language, 2, 41-63. doi: 10.1017/S030500090000088X

Maemura, Y., & Horita, M. (2012). Humour in negotiations: A pragmatic analysis of

humour in simulated negotiations. Group Decision and Negotiation, 21, 821-838.

doi:10.1007/s10726-011-9251-9

Melis, A. P., Call, J., & Tomasello, M. (2006). Chimpanzees (Pan troglodytes) conceal

visual and auditory information from others. Journal of Comparative Psychology,

120, 154-162. doi:10.1037/0735-7036.120.2.154

Norrick, N. R., & Spitz, A. (2008). Humor as a resource for mitigating conflict in

interaction. Journal of Pragmatics, 40, 1661-1686.

doi:10.1016/j.pragma.2007.12.001

182

Nottebohm, F. (1972). The origins of vocal learning. The American Naturalist, 106, 116-

140.

Nottebohm, F. (1970). Ontogeny of bird song. Science, 167(3920), 950-956.

Nottebohm, F, & Nottebohm, M. (1969). The parrots of Bush Bush. Animal Kingdom, 72,

19-23.

Penna, M., Pottstock, H., & Velasquez, N. (2005). Effect of natural and synthetic noise

on evoked vocal responses in a frog of the temperate austral forest. Animal

Behaviour, 70, 639-651.

Pepperberg, I. M. (1985). Social modeling theory: A possible framework for

understanding avian vocal learning. The Auk, 102(4), 854-864.

Pepperberg, I. M. (1994). Vocal learning in Grey parrots (Psittacus erithacus): Effects of

social interaction, reference, and context. The Auk, 111(2), 300-313.

Péron, F., Chadard, C., Nagle, L., & Bovet, D. (2011). Do African grey parrots

(Psittacus erithacus) know what a human experimenter does and does not see?

Behavioural Processes, 87, 237-240. doi:10.1016/j.beproc.2011.04.001

Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind?

Behavioral and Brain Sciences, 1, 515-526. doi:10.1017/S0140525X00076512

Schiffrin, D. (1985). Everyday argument: The organization of diversity in talk. In: D.

Tuen, & A. Van (Eds.), Handbook of discourse analysis: Discourse and dialogue,

(pp. 35-46). London: Academic Press.

Sclavi, M. (2008). The role of play and humor in creative conflict management.

Negotiation Journal, 24, 157-180. doi:10.1111/j.1571-9979.2008.00175.x

Seibert, L. M. (2006). Social behavior of Psittacine birds. In A. U. Luescher (Ed.),

183

Manual of parrot behavior (pp. 43-48). Ames: Blackwell Publishing.

Serpell, J. (1981). Duets, greetings and triumph ceremonies: Analogous displays in the

parrot genus Trichoglossus. Zeitschrift Tierpsychologie, 55, 268-283.

doi:10.1111/j.1439-0310.1981.tb01272.x

Slabbekoorn, H., & Peet, M. (2003). Birds sing at a higher pitch in urban noise. Nature,

424(6946), 267.

Stein, N. L., & Albro, E. R. (2001). The origins and nature of arguments: Studies in

conflict understanding, emotion, and negotiation. Discourse Processes, 32, 113-

133. doi:10.1080/0163853X.2001.9651594

Terrace, H. S. (1986). Nim: A chimpanzee who learned signed language. New York:

Columbia University Press.

Weiss, D. M., & Sachs, J. (1981). Persuasive strategies used by preschool children.

Discourse Processes, 14, 55-72. doi: 10.1080/01638539109544774

Wright, T. F. (1996). Regional dialects in the contact call of a parrot. Proceedings of the

Royal Society: Biological Sciences, 263, 867–872. doi:10.1098/rspb.1996.0128

Wootton, A. J. (1981). Two request forms of four year olds. Journal of Pragmatics, 5(6),

511-523.

184