Research Article Recognizing Intentions in Infant-Directed Speech Evidence for Universals Gregory A. Bryant and H. Clark Barrett University of California, Los Angeles ABSTRACT—In all languages studied to date, distinct pro- sodic contours characterize different intention categories of infant-directed (ID) speech. This vocal behavior likely exists universally as a species-typical trait, but little research has examined whether listeners can accurately recognize inten- tions in ID speech using only vocal cues, without access to semantic information. We recorded native-English-speaking mothers producing four intention categories of utterances (prohibition, approval, comfort, and attention) as both ID and adult-directed (AD) speech, and we then presented the utterances to Shuar adults (South American hunter-horti- culturalists). Shuar subjects were able to reliably distinguish ID from AD speech and were able to reliably recognize the intention categories in both types of speech, although per- formance was significantly better with ID speech. This is the first demonstration that adult listeners in an indigenous, nonindustrialized, and nonliterate culture can accurately infer intentions from both ID speech and AD speech in a language they do not speak. A major function of speech is the communication of intentions. When people speak, they form their utterances so that others will grasp their meaning. In everyday conversation between adult native language users, intentions can be conveyed through multiple channels, including the syntax and semantics of their language, but also through other means, such as prosody. Ac- curately communicating intentions to infants, however, presents a particular challenge. Because infants are not yet linguistically competent, speakers cannot make use of the full range of lan- guage systems normally available for communicating intentions between adult language speakers, including grammar and even the meanings of words themselves. The fact that adults often speak differently to infants than they do to adults is consistent with this view. In particular, when adults talk to infants, they frequently exaggerate prosodic cues normally used for conveying intentions in adult-directed (AD) speech. This pattern of exaggerated prosody is called infant- directed (ID) speech. A variety of reasons why adults use ID speech have been proposed. Likely functions include eliciting infants’ attention (Fernald & Simon, 1984; Werker & McLeod, 1989) and communicating affective intentions (Fernald, 1989, 1992). More controversial proposals include the idea that ID speech helps children learn aspects of language, such as vowel categories (Kuhl et al., 1997; Trainor & Desjardins, 2002) and grammar (Christophe, Nespor, Guasti, & Van Ooyen, 2003; Morgan & Demuth, 1996). Fernald (1992) suggested that the function of ID speech changes over development. Initially, the speech signal might serve to direct infants’ attention and mod- ulate arousal and affect, but by the 2nd year, ID speech can fulfill more specific linguistic purposes. Distinctive ID speech has been found in all languages studied to date, and it manifests itself similarly, with only subtle varia- tion (Falk, 2004; Fernald, 1992). For instance, relative to AD speech, ID speech often has overall higher mean fundamental frequency (F0), wider F0 range, more exaggerated F0 and in- tensity contours, and more musical rhythmic properties than AD speech (Fernald, 1989). Fernald (1992) described similarities in how pitch contours (i.e., F0 values represented over time) relate to communicative intentions across several languages. For ex- ample, prohibition utterances are often characterized by low F0, narrow F0 range, and staccato-like bursts. In contrast, approval vocalizations generally have high average F0, wide F0 range, and a prominent F0 rise-fall contour. These acoustic configu- rations modulate infants’ attention and subsequent behavior in expected ways without relying on verbal commands that are not readily understood. Not only are there apparent universals in production, but there is strong evidence that infants everywhere have a com- Address correspondence to Gregory A. Bryant, Department of Communication Studies, University of California, Los Angeles, 2303 Rolfe Hall, Los Angeles, CA 90095-1563, e-mail: gabryant@ucla.edu. PSYCHOLOGICAL SCIENCE 746 Volume 18—Number 8 Copyright r 2007 Association for Psychological Science