FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W1481707662> ?p ?o ?g. }

• W1481707662 abstract "Predictability effects in adult-directed and infant-directed speech: Does the listener matter? John K. Pate (j.k.pate@sms.ed.ac.uk) and Sharon Goldwater (sgwater@inf.ed.ac.uk) School of Informatics, University of Edinburgh, 10 Crichton St., Edinburgh EH8 9AB, UK Abstract A well-known effect in speech production is that more pre- dictable words tend to be phonetically reduced. Recent work has suggested that predictability effects result from hardwired properties of the language production system, rather than ac- tive modulation by the talker to accommodate the listener. However, these studies investigated only minor manipulations of listener characteristics. Here, we examine predictability ef- fects with two very different listener populations: adults and preverbal infants. Using mixed effects regressions on spon- taneous speech corpora, we compare the effect of word fre- quency, probability in context, and previous mention on word duration in adult-directed and infant-directed speech. We find that the effects of preceding context and word frequency differ according to listener. Contrary to previous work, these results suggest that talkers do modulate the phonetic effects of pre- dictability based on listener characteristics. To our knowledge, this study is also the first published analysis of predictability effects in infant-directed speech. Introduction It has long been known that more predictable words (e.g., those with higher frequency or previous mentions in the dis- course) tend to be phonetically reduced (e.g., with shorter du- ration and less distinct vowels). Recent investigations into this phenomenon have focused on an information-theoretic explanation known as the Smooth Signal hypothesis (Aylett & Turk, 2004) or the Uniform Information Density hypoth- esis (Frank & Jaeger, 2008). Under this theory, speech is viewed as a method of communicating information across a noisy channel. The most efficient encoding of information is one in which there is a constant cost (in terms of time or effort) for transmitting a single unit of information, so when words are more predictable (conveying less information), a rational speaker will tend to pronounce them more quickly (reducing time) or articulate less clearly (reducing effort). This theory assumes that speech production is optimized for efficient communication, but does not specify the source of this optimization. One possibility is that optimization is strictly offline: the talker’s speech production system is op- timized for communication across many talker-listener pairs through talker-based mechanisms such as priming, lexical ac- cess, and articulatory practice. Another possibility is that some or even most optimization occurs online, with talkers adapting their pronunciation to the immediate needs of their listeners or communicative goals, for example by reducing the pronunciation of a word only if they believe that the cur- rent listener would find it predictable. Earlier work on predictability effects often assumed online modulation of these effects in response to listeners’ needs (e.g., Lindblom, 1990), but more recent studies have cast doubt on this hypothesis. For example, Bard et al. (2000) analyzed data from dialogues between pairs of subjects nego- tiating a route on a map, where the subjects’ maps differed slightly. Bard et al. found that talkers reduced the intelligi- bility and duration of words they had previously mentioned, regardless of whether the current listener had heard or seen these words before. Bard et al. also found that talkers re- duced words that had previously been mentioned only by the listener, but they suggest that this effect could be the result of talkers forgetting who mentioned the word previously, rather than explicitly adjusting to listener knowledge. Put another way, talker-based priming mechanisms, rather than listener modeling, are sufficient to explain the result. While Bard et al.’s study suggests that talkers do not main- tain detailed models of listeners’ discourse knowledge to de- termine word predictability (and thus pronunciation), it is still possible that talkers modulate predictability effects for the benefit of the listener in response to more general knowledge about listener characteristics, such as the listener’s overall lin- guistic competence, or communicative goals. In this paper, we present evidence in favor of this weaker version of listener-based pronunciation adaptation. We do so by examining the effects of predictability on word duration in a type of speech that, to our knowledge, has not been studied previously for this purpose: infant-directed speech (IDS). By comparing predictability effects in IDS with those in adult- directed speech (ADS), we aim to tease apart two hypotheses regarding the role of the listener in determining word pre- dictability and associated reductions in pronunciation: • Hypothesis 1: All effects of predictability on pronunci- ation are hard-wired into the language production system (or are otherwise talker-based), and so are not subject to modulation in response to the listener’s needs. • Hypothesis 2: Some predictability effects are subject to modulation by the talker according to general knowledge about the listener’s needs and communicative goals. If Hypothesis 1 is correct, we would expect no differences between ADS and IDS in terms of which predictability fac- tors affect word duration, or the direction of the effects. Talk- ers may speak more slowly in IDS, or stretch all the pre- dictability effects in the same direction because communica- tion with infants represents a noisier channel. But even if IDS talkers are not interested in communicating efficiently with a preverbal infant, or assume little or no language knowledge on the part of the infant, predictability effects persist because they are results of properties of the talker alone. Under Hypothesis 2, however, we would expect some dif- ferences in predictability effects between ADS and IDS. For" @default.
• W1481707662 created "2016-06-24" @default.
• W1481707662 creator A5075564798 @default.
• W1481707662 creator A5081404974 @default.
• W1481707662 date "2011-01-01" @default.
• W1481707662 modified "2023-09-23" @default.
• W1481707662 title "Predictability effects in adult-directed and infant-directed speech: Does the listener matter?" @default.
• W1481707662 cites W1581317411 @default.
• W1481707662 cites W1801810991 @default.
• W1481707662 cites W1986424291 @default.
• W1481707662 cites W2006660745 @default.
• W1481707662 cites W2036572838 @default.
• W1481707662 cites W2071402670 @default.
• W1481707662 cites W2089654579 @default.
• W1481707662 cites W2097067028 @default.
• W1481707662 cites W2118938353 @default.
• W1481707662 cites W2140188190 @default.
• W1481707662 cites W2276277974 @default.
• W1481707662 cites W2586145851 @default.
• W1481707662 cites W2094790441 @default.
• W1481707662 hasPublicationYear "2011" @default.
• W1481707662 type Work @default.
• W1481707662 sameAs 1481707662 @default.
• W1481707662 citedByCount "0" @default.
• W1481707662 crossrefType "journal-article" @default.
• W1481707662 hasAuthorship W1481707662A5075564798 @default.
• W1481707662 hasAuthorship W1481707662A5081404974 @default.
• W1481707662 hasConcept C105795698 @default.
• W1481707662 hasConcept C138885662 @default.
• W1481707662 hasConcept C151730666 @default.
• W1481707662 hasConcept C15744967 @default.
• W1481707662 hasConcept C180747234 @default.
• W1481707662 hasConcept C197640229 @default.
• W1481707662 hasConcept C2779343474 @default.
• W1481707662 hasConcept C28490314 @default.
• W1481707662 hasConcept C33923547 @default.
• W1481707662 hasConcept C41008148 @default.
• W1481707662 hasConcept C41895202 @default.
• W1481707662 hasConcept C43617652 @default.
• W1481707662 hasConcept C86803240 @default.
• W1481707662 hasConceptScore W1481707662C105795698 @default.
• W1481707662 hasConceptScore W1481707662C138885662 @default.
• W1481707662 hasConceptScore W1481707662C151730666 @default.
• W1481707662 hasConceptScore W1481707662C15744967 @default.
• W1481707662 hasConceptScore W1481707662C180747234 @default.
• W1481707662 hasConceptScore W1481707662C197640229 @default.
• W1481707662 hasConceptScore W1481707662C2779343474 @default.
• W1481707662 hasConceptScore W1481707662C28490314 @default.
• W1481707662 hasConceptScore W1481707662C33923547 @default.
• W1481707662 hasConceptScore W1481707662C41008148 @default.
• W1481707662 hasConceptScore W1481707662C41895202 @default.
• W1481707662 hasConceptScore W1481707662C43617652 @default.
• W1481707662 hasConceptScore W1481707662C86803240 @default.
• W1481707662 hasIssue "33" @default.
• W1481707662 hasLocation W14817076621 @default.
• W1481707662 hasOpenAccess W1481707662 @default.
• W1481707662 hasPrimaryLocation W14817076621 @default.
• W1481707662 hasRelatedWork W117137278 @default.
• W1481707662 hasRelatedWork W1913141250 @default.
• W1481707662 hasRelatedWork W1979436022 @default.
• W1481707662 hasRelatedWork W1980335876 @default.
• W1481707662 hasRelatedWork W198924044 @default.
• W1481707662 hasRelatedWork W1992610688 @default.
• W1481707662 hasRelatedWork W1993049588 @default.
• W1481707662 hasRelatedWork W2003810032 @default.
• W1481707662 hasRelatedWork W2034623289 @default.
• W1481707662 hasRelatedWork W2048277749 @default.
• W1481707662 hasRelatedWork W2058787332 @default.
• W1481707662 hasRelatedWork W2069905831 @default.
• W1481707662 hasRelatedWork W2084371103 @default.
• W1481707662 hasRelatedWork W2087078059 @default.
• W1481707662 hasRelatedWork W2610616322 @default.
• W1481707662 hasRelatedWork W2611635552 @default.
• W1481707662 hasRelatedWork W3009546922 @default.
• W1481707662 hasRelatedWork W3175530665 @default.
• W1481707662 hasRelatedWork W433989144 @default.
• W1481707662 hasRelatedWork W91539319 @default.
• W1481707662 hasVolume "33" @default.
• W1481707662 isParatext "false" @default.
• W1481707662 isRetracted "false" @default.
• W1481707662 magId "1481707662" @default.
• W1481707662 workType "article" @default.