|
Article Excerpt
Contacts between the tongue and palate are needed to obstruct or constrict airflows during production of a number of speech sounds, but there is limited empirical evidence regarding the magnitude of the contact pressure between the tongue and the palate during speech. This study evaluated contact pressure generated between the tongue and alveolar ridge during consonant production. Contact pressures were measured with a transducer mounted on a palatal appliance as speakers produced stimuli that varied as a function of consonant voicing feature, manner of production, and oral-nasal coupling. Group data indicated no difference in mean articulatory contact pressure across the six consonants. However, there were differences in pressures generated across speakers. Within-speaker analysis indicated that most individuals did, in fact, produce the consonants with differing levels of contact pressures, but the ordering of the consonants based on contact pressure magnitude varied from speaker to speaker. Additional data are needed to determine the range of contact pressures to expect from nondisordered speakers and to evaluate whether there are important aerodynamic, acoustic, and perceptual outcomes of altered contact pressures. However, the current data add to the limited information on normal contact pressures generated by the tongue during speech production.
**********
The importance of tongue strength to speech sound production has been debated for a number of years. Speech is generally considered a low force activity, utilizing ~10-20% of a speaker's maximum force capabilities (see Hinton & Arkiasamy, 1997; Neary, 2004; Netsell, 1982; Thompson, Murdoch, & Stokes, 1997). As such, it has been suggested that overall tongue strength may not be a critical factor in speech sound production as long as the speaker can generate the low forces necessary for speech (Barlow & Abbs, 1983; Kent, & Rosenbek, 1987). This may be the case. However, because of limitations in the ability to index strength of tongue movements and contacts during speech as opposed to nonspeech tasks, it has been difficult to determine the expected magnitude of articulatory contact pressure (ACP) from both normal and disordered speakers; the minimally necessary ACP for perceptually normal speech; variations in contact strength as a function of the speech task or target stimulus; the impact of altered ACP on aerodynamic, acoustic, and perceptual aspects of speech; and other basic questions.
Tongue-to-palate contacts are important events in the production of consonants in many languages. To produce lingua-alveolar (LA) stops, expiratory air flow is obstructed by tongue contact against the palate, resulting in intraoral air pressure ([P.sub.o]) buildup (Subtelny, Worth, & Sakuda, 1966). Release of this [P.sub.o] creates a burst of acoustic energy that helps mark the production as a stop consonant (Stevens & Blumstein, 1978). Presumably, the ACP must be at least as high as the [P.sub.o], otherwise the articulators would be parted by the mounting [P.sub.o], affecting the ability to produce a stop burst. Likewise, for bilabial stops, lip ACP must be sufficient to contain [P.sub.o] for [TEXT NOT REPRODUCIBLE IN ASCII] and [TEXT NOT REPRODUCIBLE IN ASCII] (Hinton, 1996; Hinton & Luschei, 1992; Thompson et al., 1997). The relevance of ACP to other phonemes such as fricatives is perhaps less intuitive because there is not an associated burst. However, fricatives also require contact between articulators to create and maintain constrictions of a particular shape at various locations along the vocal tract. The [P.sub.o] builds behind the constriction, and frication noise is generated as air flows through or around the site of narrowing (Stevens, 1971).
In the past 10 years there has been a resurgence of interest in measuring the magnitude of ACP of the tongue against the palate ([ACP.sub.tongue]) as a means of indexing tongue strength during speech (Murdoch, Goozee, Veidt, Scott, & Meyers, 2004; Searl, 2003) and nonspeech tasks (e.g., McHenry, Minton, Wilson, & Post, 1994; Solomon, Robin, & Luschei, 2000; Ward, Theodoros, Murdoch, & Silburn, 2000). Clinically, a primary impetus for such work has been a long-standing question regarding the extent to which altered tongue strength impacts speech production. Reductions in tongue strength have been reported for individuals with speech disorders resulting from a variety of causes including amyotrophic lateral sclerosis (ALS) (Dworkin & Hartman, 1979; Langmore & Lehman, 1994), traumatic brain injury (Goozee, Murdoch & Theodoros, 2001; Steirwalt, Robin, Solomon, Weiss, & Max, 1995), Parkinson's disease (Solomon, Lorell, Robin, Rodnitzky, & Luschei, 1995; Ward et al., 2000), and stroke (Thompson, Murdoch, & Stokes, 1995), among others (e.g., Dworkin & Aronson, 1986; Murdoch, Attard, Ozanne, & Stokes, 1995). However, studies attempting to correlate altered tongue strength to perceptual attributes of speech, such as articulatory precision or speech intelligibility, have produced equivocal results with some finding significant relationships (e.g., Langmore & Lehman, 1994; Solomon et al., 1995) and others not (Depaul & Brooks, 1993; Solomon et al., 2000; Thompson et al., 1995).
A primary limitation in these studies of disordered speakers is that tongue strength (indexed as either force or pressure) is often evaluated during nonspeech rather than speech tasks. Correlations are then evaluated between the nonspeech measure and some aspect of speech such as intelligibility. The ability to measure contact pressures during speech would make it possible to address this long-standing issue in a more straightforward manner. That is, one could index the "strength" of articulatory activity during the event of interest, namely speech, rather than correlating nonspeech measures to speech. As Kent et al. (1987) have highlighted, nonspeech tasks (particularly maximum performance ones) do not reflect well the demands of speech production. With the ability to measure contact pressures during speech it would be possible to investigate a wide range of clinical populations for whom reduced strength of articulators is potentially at issue (e.g., ALS, paresis following stroke, etc.) to determine whether reductions in nonspeech strength translate into changes in strength of articulatory contacts that have any importance in terms of aerodynamic, acoustic, or perceptual features of the resulting speech. As a precursor to evaluation of clinical populations, however, a more complete description of ACP in nondisordered speakers is needed as a referent.
In the 1960s and early 1970s, reports of [ACP.sub.tongue] during speech appeared in the literature with some regularity although the total number of studies was small. The focus was on normal speakers and dealt mainly with phonetic issues (Brown, McGlone, & Proffit, 1973; Malecot, 1966; McGlone, & Proffit, 1967; McGlone, Proffit, & Christiansen, 1967; Proffit, Palmer, & Kydd, 1965). [ACP.sub.tongue] values for comparable stimuli have varied substantially across these studies from less than 10gm/[cm.sup.2] (Proffit et al., 1965) to ~85gm/[cm.sup.2] (Brown et al., 1973). There also is disagreement whether phonemes can be distinguished by peak ACP measures (Brown et al., 1973; Malecot, 1966; McGlone & Proffit, 1967; McGlone et al., 1967; Proffit et al., 1965). Interspeaker variability in [ACP.sub.tongue] has been large (Malecott, 1966; McGlone & Proffit, 1967; Proffit et al., 1965), but the degree of expected intraspeaker variability is unclear, with the investigations of the matter in disagreement (Malecot, 1966; McGlone & Proffit, 1967).
Investigators have identified several limitations to these early studies of ACP including small subject pools, a restricted set of phonemes considered, limited types of syllable constructions, and variable placement of transducers (Hinton & Luschei, 1992; Murdoch et al., 2004; Searl, 2003). Perhaps the greatest limitation was the technology available at the time. Concerns regarding physical dimensions and response characteristics of early transducers have been reviewed by Hinton and Luschei (1992), Murdoch et al. (2004) and Searl (2003). Briefly, some of the transducers were large enough to disrupt normal speech production as several authors reported anecdotally (e.g., McGlone & Proffit, 1967; McGlone et al., 1967; Proffit et al., 1965). The response characteristics of these early ACP sensing instruments also have been questioned, with particular concerns about linearity and temperature compensation features. Limitations in the available technology may explain some of the discrepant results among the earlier works, although small subject pools in conjunction with marked between speaker variability in contact pressure values are also likely contributors.
Advances in technology over the past several decades have brought about the opportunity to reassess [ACP.sub.tongue] with transducers that may be more suited for the task. Hinton and Luschei (1992) evaluated a miniature transducer (Entran EPL-2001-10) that they determined had...
|
