Word recognition from acoustic onsets and acoustic offsets: Effects of cohort size and syllabic stress

Abstract

In the traditional gating technique, subjects hear increasing amounts of word-onset information from spoken words until the words can be correctly identified. The experiment reported here contrasted word-onset gating with results when words were gated from their word endings. A significant recognition advantage for words gated from their onsets was demonstrated. This effect was eliminated, however, when we took into account the number of word possibilities that shared overlapping phonology and the same stress pattern as the target words at their recognition points. These results support the position that the perceptual advantage of word-initial information can be understood within a general goodness-of-fit model of spoken word recognition.A considerable amount of evidence is now available to show that spoken words can be correctly identified when far less than their full acoustic duration has been heard. This can be demonstrated by using the word-onset "gating" technique first developed by Grosjean (1980;Cotton & Grosjean, 1984). In this technique, a listener is allowed to hear increasing amounts of word-onset information until the word can be correctly identified: for example, the subject hears the first 50 ms of a word, then the first 100 ms of the word, then the first 150 ms of the word, and so on, until the word is correctly reported. In fluent discourse, words are often underarticulated (Lieberman, 1963;Pickett & Pollack, 1963;Wingfield, Alexander, & Cavigelli, 1994), and the word-onset duration necessary for identification will vary with the clarity of the speech signal (Grosjean, 1985;Nooteboom & Doodeman, 1984). It is nevertheless the case that words heard in isolation (i.e., without a constraining sentence context) can be recognized, on average, with little more than the first half of their full acoustic duration (Grosjean, 1980;Marslen-Wilson, 1984;Tyler, 1984).

Testing the protracted lexical restructuring hypothesis: The effects of position and acoustic–phonetic clarity on sensitivity to mispronunciations in children and adults

“…In fact, overreliance on word-initial information would produce serious comprehension difficulties whenever word-initial errors are made by either the speaker or the listener. Although word-initial information may be given priority, listeners can recognize words on the basis of partial information even in backward gating tasks (Salasoo & Pisoni, 1985;Walley, 1988;Wingfield et al, 1997). More recent versions of the cohort model acknowledge these difficulties (Gaskell & Marslen-Wilson, 1999;Marslen-Wilson, 1987).…”

“…Adults can recognize words on the basis of partial word-initial or even word-final information (Grosjean, 1980;Marslen-Wilson, 1987;Marslen-Wilson & Welsh, 1978;Salasoo & Pisoni, 1985;Walley, 1988;Wingfield, Goodglass, & Lindfield, 1997), suggesting that word recognition proceeds on the basis of matching partially analyzed input with segmentally organized lexical representations in memory. This phenomenon most typically has been studied using a gating task in which participants hear successively longer fragments of words, for instance, beginning with the initial (or final) 50 ms and increasing by small increments until participants can reliably identify the words without any subsequent change in response.…”

“…This phenomenon most typically has been studied using a gating task in which participants hear successively longer fragments of words, for instance, beginning with the initial (or final) 50 ms and increasing by small increments until participants can reliably identify the words without any subsequent change in response. Adults require less word-initial than word-final information to recognize words (Salasoo & Pisoni, 1985;Walley, 1988;Wingfield et al, 1997). Furthermore, words with sparse onset neighborhoods are recognized faster than those with dense onset neighborhoods, even when the overall neighborhood density of the two sets of words is equivalent (Vitevitch, 2002).…”

Manulex-infra: Distributional characteristics of grapheme—phoneme mappings, and infralexical and lexical units in child-directed written material

“…Finally, findings suggest that word recognition in adults is also partially determined by the location of the uniqueness point (Kwantes & Mewhort, 1999;Lindell, Nicholls, Kwantes, & Castles, 2005; but see Lamberts, 2005). This variable, initially manipulated in auditory word-recognition studies (e.g., Marslen-Wilson, 1984;Radeau & Morais, 1990;Wingfield, Goodglass, & Lindfield, 1997), is defined as the point (phoneme) in the target word where only one lexical candidate remains, given the sequential property of speech. Transposed to reading, the orthographic point of uniqueness corresponds to the serial position of the letter in the word where only one lexical candidate remains, considering the letter strings serially from left to right.…”

Effects of stress typicality during spoken word recognition by native and nonnative speakers of English: Evidence from onset gating

“…In her discussion of the potential importance of wordinitial information during spoken word recognition, Walley (1988) noted that word-initial information reaches the listener first (especially pertinent in view of the transience of the speech waveform) and that word-initial information is more resistant to change by phonological processes. Indeed, studies in which a variety of paradigms were used, including noise replacement /addition and gating, have shown that word-initial information facilitates spoken word recognition more than does word-final information (e.g., Cole & Jakimik, 1980;Nooteboom, 1981;Salasoo & Pisoni, 1985;Walley, 1988;Wingfield, Goodglass, & Lindfield, 1997).…”