Synthesizing meaning and processing approaches to prosody: performance matters

Metric Structure and Rhyme Predictability Modulate Speech Intensity During Child-Directed and Read-Alone Productions of Children's Literature

Temporal and phonological predictability in children's literature may support early literacy acquisition. Realization of predictive structure in caregiver prosody could guide children's attention during shared reading, thereby supporting reading subskill development. However, little is known about how predictive structure is realized prosodically during child-directed reading. We investigated whether speakers use word intensity to signal predictive metric and rhyme structure in child-directed and read-alone productions of The Cat in the Hat (Dr. Seuss, 1957), by modeling maximum intensity (dB) of monosyllabic words as a function of metric strength, rhyme predictability, and a set of control parameters. In the control model, intensity increased with lower lexical frequency, capitalization, first mention, and likelihood of a syntactic boundary. Metric structure predicted word intensity beyond these control factors in a hierarchical manner: words aligned with beat one in a 6/8 metric structure were produced with highest intensity, words aligned with beat four were produced with intermediate intensity, and words aligned with all other beats were produced with the lowest intensity. Additionally, phonologically predictable rhyme targets were reduced in intensity. The effects of meter and rhyme were not moderated by the presence of a child audience. These results demonstrate that predictability along multiple dimensions is encoded during reading of poetic children's literature, and that metric structure is realized hierarchically in word intensity. Further, the manner by which predictability is encoded in word intensity differs from that previously reported for word duration in this corpus (Breen, 2018), demonstrating that intensity and duration present nonidentical prosodic information channels.

Some inferences still take time: Prosody, predictability, and the speed of scalar implicatures

Experimental pragmatics has gained many insights from understanding how people use weak scalar terms (like some) to infer that a stronger alternative (like all) is false. Early studies found that comprehenders initially interpret some without an upper bound, but later results suggest that this inference is sometimes immediate (e.g., Grodner, Klein, Carbary, & Tanenhaus, 2010). The present paper explores whether rapid inferencing depends on the prosody (i.e., summa rather than some of) or predictability of referring expressions (e.g., consistently using some to describe subsets). Eye-tracking experiments examined looks to subsets (2-of-4 socks) and total sets (3-of-3 soccer balls) following some and found early preferences for subsets in predictable contexts but not in less predictable contexts (Experiment 1 and 2). In contrast, there was no reliable prosody effect on inferencing. Changes in predictability did not affect judgments of the naturalness of some, when a discourse context was available (Experiment 3). However, predictable contexts reduced variability in speakers' descriptions of subsets and total sets (Experiment 4). Together, these results demonstrate that scalar inferences are often delayed during comprehension, but reference restriction is rapid when set descriptions can be formulated beforehand.

Variation in the speech signal as a window into the cognitive architecture of language production

The pronunciation of words is highly variable. This variation provides crucial information about the cognitive architecture of the language production system. This review summarizes key empirical findings about variation phenomena, integrating corpus, acoustic, articulatory, and chronometric data from phonetic and psycholinguistic studies. It examines how these data constrain our current understanding of word production processes and highlights major challenges and open issues that should be addressed in future research.

The role of linguistic experience in the processing of probabilistic information in production

Speakers track the probability that a word will occur in a particular context and utilize this information during phonetic processing. For example, content words that have high probability within a discourse tend to be realized with reduced acoustic/articulatory properties. Such probabilistic information may influence L1 and L2 speech processing in distinct ways (reflecting differences in linguistic experience across groups and the overall difficulty of L2 speech processing). To examine this issue, L1 and L2 speakers performed a referential communication task, describing sequences of simple actions. The two groups of speakers showed similar effects of discourse-dependent probabilistic information on production, suggesting that L2 speakers can successfully track discourse-dependent probabilities and use such information to modulate phonetic processing.

Using prosody to infer discourse prominence in cochlear-implant users and normal-hearing listeners

Cochlear implants (CIs) provide speech perception to adults with severe-to-profound hearing loss, but the acoustic signal remains severely degraded. Limited access to pitch cues is thought to decrease sensitivity to prosody in CI users, but co-occurring changes in intensity and duration may provide redundant cues. The current study investigates how listeners use these cues to infer discourse prominence. CI users and normal-hearing (NH) listeners were presented with sentences varying in prosody (accented vs. unaccented words) while their eye-movements were measured to referents varying in discourse status (given vs. new categories). In Experiment 1, all listeners inferred prominence when prosody on nouns distinguished categories ("SANDWICH"→not sandals). In Experiment 2, CI users and NH listeners presented with natural speech inferred prominence when prosody on adjectives implied contrast across both categories and properties ("PINK horse"→not the orange horse). In contrast, NH listeners presented with simulated CI (vocoded) speech were sensitive to acoustic differences in prosody, but did not use these cues to infer discourse status. Together, this suggests that exploiting redundant cues for comprehension varies with the demands of language processing and prior experience with the degraded signal.

Explicit and Emergent Mechanisms of Information Status

It is well established that language production and comprehension are influenced by information status, for example, whether information is given, new, topical, or predictable, and many scholars suggest that an important component of information status is keeping track of what information is in common ground (i.e., what is shared), and what is not. Information status affects both speakers' choices (e.g., word order, pronoun use, prosodic prominence) and how listeners interpret the speaker's meaning (e.g., Chafe, 1994; Prince, 1981). Although there is a wealth of scholarly work on information status (for a review, see Arnold, Kaiser, Kahn, & Kim, 2013), there is no consensus on the mechanisms by which it is used, and in fact relatively little discussion of the underlying representations and psycholinguistic mechanisms. Moreover, a major challenge to understanding information status is that its effects are notoriously variable. This study considers existing proposals about information status, focusing on two questions: (a) how is it represented; and (b) by what mechanisms is it used? I propose that it is important to consider whether representations and mechanisms can be classified as either explicit or emergent. Based on a review of existing evidence, I argue that information status representations are most likely emergent, but the mechanisms by which they are used are both explicit and emergent. This review provides one of the first considerations of information status processing across multiple domains.

Dynamically adapted context-specific hyper-articulation: Feedback from interlocutors affects speakers' subsequent pronunciations

We ask whether speakers can adapt their productions when feedback from their interlocutors suggests that previous productions were perceptually confusable. To address this question, we use a novel web-based task-oriented paradigm for speech recording, in which participants produce instructions towards a (simulated) partner with naturalistic response times. We manipulate (1) whether a target word with a voiceless plosive (e.g., pill) occurs in the presence of a voiced competitor (bill) or an unrelated word (food) and (2) whether or not the simulated partner occasionally misunderstands the target word. Speakers hyper-articulated the target word when a voiced competitor was present. Moreover, the size of the hyper-articulation effect was nearly doubled when partners occasionally misunderstood the instruction. A novel type of distributional analysis further suggests that hyper-articulation did not change the target of production, but rather reduced the probability of perceptually ambiguous or confusable productions. These results were obtained in the absence of explicit clarification requests, and persisted across words and over trials. Our findings suggest that speakers adapt their pronunciations based on the perceived communicative success of their previous productions in the current environment. We discuss why speakers make adaptive changes to their speech and what mechanisms might underlie speakers' ability to do so.

The (in)dependence of articulation and lexical planning during isolated word production

The number of phonological neighbors to a word (PND) can affect its lexical planning and pronunciation. Similar parallel effects on planning and articulation have been observed for other lexical variables, such as a word's contextual predictability. Such parallelism is frequently taken to indicate that effects on articulation are mediated by effects on the time course of lexical planning. We test this mediation assumption for PND and find it unsupported. In a picture naming experiment, we measure speech onset latencies (planning), word durations, and vowel dispersion (articulation). We find that PND predicts both latencies and durations. Further, latencies predict durations. However, the effects of PND and latency on duration are independent: parallel effects do not imply mediation. We discuss the consequences for accounts of lexical planning, articulation, and the link between them. In particular, our results suggest that ease of planning does not explain effects of PND on articulation.

Reduction in Prosodic Prominence Predicts Speakers' Recall: Implications for Theories of Prosody

Repeated words are often reduced in prosodic prominence, but the underlying mechanisms remain unclear. The present study contrasted two theories: does prosodic reduction reflect the choice of a particular linguistic form, or does ease of retrieval within the language production system lead to facilitated, less prominent productions? One test of facilitation-based theories is suggested by findings on human memory: Whether a second presentation of an item benefits later memory is predicted by the item's availability at the time of the second presentation. If prosodic reduction partially reflects facilitated retrieval, it should predict later memory. One naïve participant described to another participant routes on a map. Critical items were mentioned twice. Following the map task, the speaker attempted written recall of the mentioned items. As expected, acoustic intensity of the second mentions predicted later recall in the same way that difficulty of retrieval has in other tasks. This pattern suggests that one source of prosodic reduction is facilitation within the language production system.