Preaspiration in the Nordic Languages: Synchronic and Diachronic ...
Preaspiration in the Nordic Languages: Synchronic and Diachronic ... Preaspiration in the Nordic Languages: Synchronic and Diachronic ...
ation, voiceless nasality before fortis stops may have been a feature employed by some (but not all) speakers as one of the correlates of the production of sequences in which a fortis stop follows a nasal. Again, in some areas—notably only the areas in which preaspiration becomes the norm—the structure of phonetic variation changes, and voiced nasals cease to be contained within the norm. As a result, voiceless or partially voiceless nasals become the norm. In this account, we also have to assume that, unlike preaspiration, this alleged non-normative PN voicelessness in nasals has ceased to be part of the norm in the production of such sequences in CSw. As before, no speaker is at any time required to make any phonological reinterpretation. Instead, the limits of phonetic variation in the expression of these sequences becomes reduced, as voiced variants become less common. The development of postaspiration instead of voiceless nasality, which we find in Northern Icelandic and the Western Åland dialects, can be accommodated in a similar scenario. Changes involving nasal + stop sequences in other languages may be treated as E/C changes as well. In Bantu languages, reconstruction suggests that Proto-Bantu had word-initial sequences of nasal and voiceless, unaspirated stops. These are reflected in various ways in the modern descendants of Proto-Bantu. For example, *nt can yield, for instance, [nt], [t], [t] or [t], depending on language and context. (cf. e.g. Hinnebusch 1989, Huffman & Hinnebusch 1998). As in the Nordic languages, these changes can be accommodated in a model that operates with the restructuring of phonetic variation rather than phonological reinterpretation. The basic assumption is that both a tendency for early voice offset (causing voicelessness in the nasal) and a tendency for late voice onset (causing postaspiration) were present as non-normative variants in the production of these sequences. This could explain why the eventual outcome varies so much in different languages and contexts. 5.5 Summary Let us briefly recapitulate the scenario of the emergence of preaspiration suggested here. First, it was suggested that the ultimate origins of preaspirations have to do with vocal fold physiology, i.e. that a certain type of larynx is more prone to causing anticipatory voice offset than are – 238 –
others. In this way, some speakers may produce “fortuitous” preaspirations that are short but salient. Further, it was suggested that a process of change similar to that proposed by Ohala (1981, 1993) could account for the introduction of preaspirated variants into speech production. Based on the data on phonetic variation in CSw, it was suggested that such preaspirated variants may not have been employed by all speakers, but were still a part of the pronunciation norms. Thus, preaspiration was nonnormative, i.e. the timing of voice offset relative to stop closure in the production of fortis stops was—and still is in most of Scandinavia— relatively free. The time at which this relative freedom in voice offset became established is unknown, but in the scenario proposed here one must assume that it was already the dominant pattern in PN. At this stage, then, PN had the phonetic preconditions necessary for preaspiration to become normative. This did occur in several areas. In each area, the exact conditions under which the transition from nonnormative to normative status differed, and consequently the phonological distribution of preaspiration differs considerably between dialects. Also, in some areas a typologically more usual development occurred and postaspirations became normative. The change from non-normative to normative preaspiration, it was proposed, involves nothing more than the restructuring of phonetic variation such that variants without preaspiration disappear from the pool of variation. Consequently, the relative freedom in the timing of the laryngeal and oral gestures that had previously been dominant becomes more limited. 5.6 Final remarks In Ohala’s (1993:244) view, changes in norms do not, as such, constitute sound change, precisely because “they do not necessarily involve a change in pronunciation norm.” This point cannot be disputed, since “sound change” is generally used to refer to phonological change rather than phonetic change. However, in the E/C model, phonetic variation is dynamic and partly governed by convention, and thus only partly a predictable and fortuitous by-product of speech production. The data reviewed in Chapter 4 clearly demonstrate this. Changes in phonetic variation, for example an expansion or contraction in a given phonetic – 239 –
- Page 200 and 201: Table 4-16. The findings of Indrið
- Page 202 and 203: these speakers will henceforth be r
- Page 204 and 205: lenes are often produced with a voi
- Page 206 and 207: Nasal + stop sequences in the Weste
- Page 208 and 209: inte ‘not’ without the final vo
- Page 210 and 211: ange for the remaining subjects is
- Page 212 and 213: For word-medial contexts data are a
- Page 214 and 215: instances may have preaspiration on
- Page 216 and 217: 4.6 Summary and conclusions The sur
- Page 218 and 219: that the ON word-initial lenes have
- Page 220 and 221: dialect in the Åland archipelago.
- Page 222 and 223: clear that ON m, n + p, t, k sequen
- Page 224 and 225: gradual process rather than a leap
- Page 226 and 227: stops that we find in, for example,
- Page 228 and 229: oth found in V()C syllables, but ar
- Page 230 and 231: Some Icelandic linguists seem to ha
- Page 232 and 233: On a quite different note, Liberman
- Page 234 and 235: dialect. With this in mind, Hansson
- Page 236 and 237: (1997:114f), Iceland, the Faroes, J
- Page 238 and 239: observed phonological distribution
- Page 240 and 241: sense, preaspiration is linked more
- Page 242 and 243: higher degree of spectral tilt than
- Page 244 and 245: a failure to apply a process of per
- Page 246 and 247: similar to that in CSw today. Some
- Page 248 and 249: from t1 through t8, the productions
- Page 252 and 253: parameter such as VOT or F0, may th
- Page 254 and 255: identity and paralinguistic informa
- Page 256 and 257: —. 1974. “On the influence of N
- Page 258 and 259: Gillies, William. 1993. “Scottish
- Page 260 and 261: Johanson, Lars. 1998. “The histor
- Page 262 and 263: Lyttkens, Ivar Adolf & Fredrik Amad
- Page 264 and 265: —. 1995b. “Speaking rate, VOT a
- Page 266 and 267: Stölten, Katrin. 2002. Dialektalit
- Page 269: Department of Linguistics, Stockhol
ation, voiceless nasality before fortis stops may have been a feature<br />
employed by some (but not all) speakers as one of <strong>the</strong> correlates of <strong>the</strong><br />
production of sequences <strong>in</strong> which a fortis stop follows a nasal. Aga<strong>in</strong>, <strong>in</strong><br />
some areas—notably only <strong>the</strong> areas <strong>in</strong> which preaspiration becomes <strong>the</strong><br />
norm—<strong>the</strong> structure of phonetic variation changes, <strong>and</strong> voiced nasals<br />
cease to be conta<strong>in</strong>ed with<strong>in</strong> <strong>the</strong> norm. As a result, voiceless or partially<br />
voiceless nasals become <strong>the</strong> norm. In this account, we also have to<br />
assume that, unlike preaspiration, this alleged non-normative PN voicelessness<br />
<strong>in</strong> nasals has ceased to be part of <strong>the</strong> norm <strong>in</strong> <strong>the</strong> production of<br />
such sequences <strong>in</strong> CSw. As before, no speaker is at any time required to<br />
make any phonological re<strong>in</strong>terpretation. Instead, <strong>the</strong> limits of phonetic<br />
variation <strong>in</strong> <strong>the</strong> expression of <strong>the</strong>se sequences becomes reduced, as<br />
voiced variants become less common. The development of postaspiration<br />
<strong>in</strong>stead of voiceless nasality, which we f<strong>in</strong>d <strong>in</strong> Nor<strong>the</strong>rn Icel<strong>and</strong>ic <strong>and</strong> <strong>the</strong><br />
Western Ål<strong>and</strong> dialects, can be accommodated <strong>in</strong> a similar scenario.<br />
Changes <strong>in</strong>volv<strong>in</strong>g nasal + stop sequences <strong>in</strong> o<strong>the</strong>r languages may be<br />
treated as E/C changes as well. In Bantu languages, reconstruction suggests<br />
that Proto-Bantu had word-<strong>in</strong>itial sequences of nasal <strong>and</strong> voiceless,<br />
unaspirated stops. These are reflected <strong>in</strong> various ways <strong>in</strong> <strong>the</strong> modern descendants<br />
of Proto-Bantu. For example, *nt can yield, for <strong>in</strong>stance, [nt],<br />
[t], [t] or [t], depend<strong>in</strong>g on language <strong>and</strong> context. (cf. e.g. H<strong>in</strong>nebusch<br />
1989, Huffman & H<strong>in</strong>nebusch 1998). As <strong>in</strong> <strong>the</strong> <strong>Nordic</strong> languages, <strong>the</strong>se<br />
changes can be accommodated <strong>in</strong> a model that operates with <strong>the</strong> restructur<strong>in</strong>g<br />
of phonetic variation ra<strong>the</strong>r than phonological re<strong>in</strong>terpretation.<br />
The basic assumption is that both a tendency for early voice offset<br />
(caus<strong>in</strong>g voicelessness <strong>in</strong> <strong>the</strong> nasal) <strong>and</strong> a tendency for late voice onset<br />
(caus<strong>in</strong>g postaspiration) were present as non-normative variants <strong>in</strong> <strong>the</strong><br />
production of <strong>the</strong>se sequences. This could expla<strong>in</strong> why <strong>the</strong> eventual<br />
outcome varies so much <strong>in</strong> different languages <strong>and</strong> contexts.<br />
5.5 Summary<br />
Let us briefly recapitulate <strong>the</strong> scenario of <strong>the</strong> emergence of preaspiration<br />
suggested here. First, it was suggested that <strong>the</strong> ultimate orig<strong>in</strong>s of<br />
preaspirations have to do with vocal fold physiology, i.e. that a certa<strong>in</strong><br />
type of larynx is more prone to caus<strong>in</strong>g anticipatory voice offset than are<br />
– 238 –