Speech perception and phonemic restorations

Abstract

When a speech sound in a sentence is replaced completely by an extraneous sound (such as a cough or tone), the listener restores the missing sound on the bases of both prior and subsequent context. This illusory effect, called phonemic restoration (PhR), causes the physically absent phoneme to seem as real as the speech sounds which are present. The extraneous sound seems to occur along with other phonemes without interfering with their clarity. But if a silent gap (rather than an extraneous sound) replaces the same phoneme, the interruption in the sentence is more readily localized in its true position and PhRs occurs less frequently. Quantitative measures were taken both of the incidence of PhRs and of the direction and extent of temporal mislocalizations of interruptions for several related stimuli under a variety of experimental conditions. The results were connected with other auditory illusions and temporal confusions reported in the literature, and suggestions were made concerning mechanisms employed normally for verbal organization.

Perceptual restoration of a “missing” speech sound: Auditory induction or illusion?

“…injhe speech signal and because the same stimuh were repeated over and over in the course of the experiment. As to the problem of repetition itself, there is no evidence in the literature that repeated presentation reduces phonemic restoration (see Samuel, 1991;Warren & Obusek, 1971).…”

“…One of the best known and most interesting findings in speech perception research is the "phonemic restoration illusion" first demonstrated by Warren (1970;Warren & Obusek, 1971;Warren & Sherman, 1974). In Warren's original experiments, subjects listened to a sentence in which the acoustic signal pertaining to one phoneme, the fricative lsI, had been excised and replaced with an extraneous sound or with silence.…”

Auditory pattern recognition by untrained listeners

“…There is recent evidence that in speech as well, pattern recognition precedes constituent analysis, so that identification of a syllable precedes identification of its component sounds (Savin & Bever, 1970;Warren, 1971). Observations with the phonemic restoration effect also indicate that phonemes are inferred, and not perceived directly (Warren & Obusek, 1971). If, indeed, much of what passes for perception of temporal order requires prior pattern recognition, perhaps formation of recognizable groupings is a fundamental process which *This study was supported by grants from the National Science Foundation (GB-26459 and GB·36986JO, and from the University of Wisconsin-Milwaukee Graduate School.…”

Why not model spoken word recognition instead of phoneme monitoring?

“…In addition, the listener should be unable to tell where in a word the noise had occurred. It was found that this "phonemic restoration" did indeed occur: When told that a speech sound in a sentence had been replaced by noise, listeners could identify neither which speech sound was absent nor where the noise had occurred even after listening to the sentence several times (Warren 1970;Warren & Obusek 1971;Warren & Warren 1970; see also Warren 1999).…”

“…I surmised that identification of phonetic components in running speech could not be accomplished directly, but was inferred following linguistic organization. This led to a formal experiment carried out that year that demonstrated that phonetic targets required more time for identification than did the monosyllables containing them, whether the monosyllable was a word in a sentence, a word in a list, or an item in a list of nonsense syllables (Warren 1971). Further, I reported in this paper that when sentences were used, context that accelerated or delayed the identification time for a word produced a corresponding change in the time required for identification of its phonetic components.…”