Meghan Sumner

The socially weighted encoding of spoken words: a dual-route approach to speech perception

Spoken words are highly variable. A single word may never be uttered the same way twice. As listeners, we regularly encounter speakers of different ages, genders, and accents, increasing the amount of variation we face. How listeners understand spoken words as quickly and adeptly as they do despite this variation remains an issue central to linguistic theory. We propose that learned acoustic patterns are mapped simultaneously to linguistic representations and to social representations. In doing so, we illuminate a paradox that results in the literature from, we argue, the focus on representations and the peripheral treatment of word-level phonetic variation. We consider phonetic variation more fully and highlight a growing body of work that is problematic for current theory: words with different pronunciation variants are recognized equally well in immediate processing tasks, while an atypical, infrequent, but socially idealized form is remembered better in the long-term. We suggest that the perception of spoken words is socially weighted, resulting in sparse, but high-resolution clusters of socially idealized episodes that are robust in immediate processing and are more strongly encoded, predicting memory inequality. Our proposal includes a dual-route approach to speech perception in which listeners map acoustic patterns in speech to linguistic and social representations in tandem. This approach makes novel predictions about the extraction of information from the speech signal, and provides a framework with which we can ask new questions. We propose that language comprehension, broadly, results from the integration of both linguistic and social information.

The Role of Variation in the Perception of Accented Speech.

Phonetic variation has been considered a barrier that listeners must overcome in speech perception, but has been proved beneficial in category learning. In this paper, I show that listeners use within-speaker variation to accommodate gross categorical variation. Within the perceptual learning paradigm, listeners are exposed to p-initial words in English produced by a native speaker of French. Critically, listeners are trained on these words with either invariant or highly-variable VOTs. While a gross boundary shift is made for participants exposed to the variable VOTs, no such shift is observed after exposure to the invariant stimuli. These data suggest that increasing variation improves the mapping of perceptually mismatched stimuli.

Lexical Inhibition and Sublexical Facilitation Are Surprisingly Long Lasting.

When a listener hears a word (beef), current theories of spoken word recognition posit the activation of both lexical (beef) and sublexical (/b/, /i/, /f/) representations. No lexical representation can be settled on for an unfamiliar utterance (peef). The authors examined the perception of nonwords (peef) as a function of words or nonwords heard 10-20 min earlier. In lexical decision, nonword recognition responses were delayed if a similar word had been heard earlier. In contrast, nonword processing was facilitated by the earlier presentation of a similar nonword (baff-paff). This pattern was observed for both word-initial (beef-peef), and word-final (job-jop) deviation. With the word-in-noise task, real word primes (beef) increased real word intrusions for the target nonword (peef), but only consonant-vowel (CV) or vowel-consonant (VC) intrusions were increased with similar pseudoword primes (baff-paff). The results across tasks and experiments support both a lexical neighborhood view of activation and sublexical representations based on chunks larger than individual phonemes (CV or VC sequences).

The social weight of spoken words

Speech serves a linguistic function, cueing sounds and words, and a social function, cueing talkers and their social attributes. Listeners readily map sound patterns in speech to social representations. This mapping introduces social biases on the recognition and encoding of sound patterns produced by different groups and individuals.

Effects of phonetically-cued talker variation on semantic encoding

This study reports equivalence in recognition for variable productions of spoken words that differ greatly in frequency. General American (GA) listeners participated in either a semantic priming or a false-memory task, each with three talkers with different accents: GA, New York City (NYC), and Southern Standard British English (BE). GA/BE induced strong semantic priming and low false recall rates. NYC induced no semantic priming but high false recall rates. These results challenge current theory and illuminate encoding-based differences sensitive to phonetically-cued talker variation. The findings highlight the central role of phonetic variation in the spoken word recognition process.

A phonetic explanation of pronunciation variant effects

Effects of word-level phonetic variation on the recognition of words with different pronunciation variants (e.g., center produced with/(out) [t]) are investigated via the semantic- and pseudoword-priming paradigms. A bias favoring clearly articulated words with canonical variants ([nt]) is found. By reducing the bias, words with different variants show robust and equivalent lexical activation. The equivalence of different word forms highlights a snag for frequency-based theories of lexical access: How are words and word productions with vastly different frequencies recognized equally well by listeners? A process-based account is proposed, suggesting that careful speech induces bottom-up processing and casual speech induces top-down processing.

Beyond lexical meaning: The effect of emotional prosody on spoken word recognition

This study employs an auditory-visual associative priming paradigm to test whether non-emotional words uttered in emotional prosody (e.g., pineapple spoken in angry prosody or happy prosody) facilitate recognition of semantically emotional words (e.g., mad, upset or smile, joy). The results show an affective priming effect between emotional prosody and emotional words independent of lexical carriers of the prosody. Learned acoustic patterns in speech (e.g., emotional prosody) map directly to social concepts and representations, and this social information influences the spoken word recognition process.

Indexical variation affects semantic spread

The role of indexical variation in spoken word recognition is constrained to acoustically rich lexical representations. Theoretically, lexical activation depends on indexical variation, but subsequent processes like associative semantic spread depend on activation strength, not indexical variation. Social psychological theories view indexical variation as integral to online processes such as persona construal. Therefore, information gleaned from indexical variation might pervade spoken word recognition more broadly. We investigate the effects of indexical variation on semantic activation in word-association and semantic-priming paradigms. Across three studies, we show that top associate responses depend on the voice of the probe word (“space” in man’s voice: time; woman’s voice: star; child’s voice: planet). Voice also affects response frequency distributions: the man’s voice receives a wider variety of weaker responses, while the woman’s and child’s voices receive fewer, stronger, responses. We also find...

Between- and Within-Speaker Effects of Bilingualism on F0 Variation.

To what extent is prosody shaped by cultural and social factors? Existing research has shown that an individual bilingual speaker exhibits differences in framing, ideology, and personality when speaking their two languages. To understand whether these differences extend to prosody we study F0 variation in a corpus of interviews with German-Italian and German-French bilingual speakers. We find two primary effects. First, a betweenspeaker effect: these two groups of bilinguals make different use of F0 even when they are all speaking German. Second, a within-speaker effect: bilinguals use F0 differently depending on which language they are speaking, differences that are consistent across speakers. These effects are modulated strongly by gender, suggesting that language-specific social positioning may play a central role. These results have important implications for our understanding of bilingualism and cross-cultural linguistic difference in general. Prosody appears to be a moving target rather than a stable feature, as speakers use prosodic variation to position themselves on cultural and social axes like linguistic context and gender.

Differences in the recognition of careful and casual speech

Previous work in spoken word recognition and speech perception has shown two seemingly conflicting patterns. While some studies have shown a processing benefit for more frequent word variants (i.e., in a casual speech mode), others have found a benefit for more canonical word forms (i.e., in a careful speech mode). This study aims to reconcile these findings, proposing that different types of processing apply to each speech mode–top-down processing for casual speech, and bottom-up for careful speech. Listeners in an auditory priming task heard natural (non-spliced) sentences spoken in either a careful or casual speech mode. The final word of the auditory prime was either semantically predictable from the preceding sentence context or unpredictable. After the audio prime, listeners responded in a lexical decision task to a visual probe: either the final word heard in the prime, an unrelated word, or a nonword. Preliminary results suggest that, regardless of speech style, reaction times are faster for related targets in the semantically predictable conditions than for unrelated targets. Crucially, responses to the target word in the careful condition are delayed compared to casual speech for semantically unpredictable sentences. The implications for the apparent paradox in previous results will be discussed.