Academia.eduAcademia.edu

Figure 7 – uploaded by Catherine Ringen

See full PDFdownloadDownload figure
(7) Spectrogram of MP’s production of the word byget ‘built (sup.)’.

Figure 7 (7) Spectrogram of MP’s production of the word byget ‘built (sup.)’.

Related Figures (13)

(1) Spectrogram of MP’s production of the word dagg ‘dew’. (2) Spectrogram of MP’s production of the word tabbe ‘mistake’
(1) Spectrogram of MP’s production of the word dagg ‘dew’. (2) Spectrogram of MP’s production of the word tabbe ‘mistake’
(5) Spectrogram of MP’s production of the word ddck ‘deck’.
(5) Spectrogram of MP’s production of the word ddck ‘deck’.
Word-final clusters with fortis stops were also examined. An example is given in (7). These were divided into two categories. First, there were clusters that can be derived from /pt/ and /kt/ sequences, in words such as képt ‘bought’ (with a short vowel; supine of képa ‘buy’) or ldkt ‘healed’ (with a long vowel; supine of /dka ‘heal’). Second, there were clusters that can be derived from /gt/ sequences, in words such as byget ‘built’ (with a short vowel; supine of bygga ‘build’) or vdgt ‘weighed’ (with a long vowel; supine of vdga ‘weigh’). Phonetically, the two types of clusters were found to be very similar. In both types, the two stop components were released and voiceless, and the cluster tended to be preceded by a slight preaspiration. The mean ei? acini i A A total of 48 intervocalic fortis clusters were recorded, consisting of the sequences [pt] and [kt] (24 of each). These clusters were invariably produced as voiceless, with both stops released. Like simple intervocalic fortis stops, they were generally preceded by a slight preaspiration. The mean duration of this preaspiration was 31 ms. Also, mean postaspiration duration was short, 26 ms, which is not indicative of any significant postaspiration percept. An example of an intervocalic fortis cluster in képte < k6/pd/e “bought (past)’ is given in (6).
Word-final clusters with fortis stops were also examined. An example is given in (7). These were divided into two categories. First, there were clusters that can be derived from /pt/ and /kt/ sequences, in words such as képt ‘bought’ (with a short vowel; supine of képa ‘buy’) or ldkt ‘healed’ (with a long vowel; supine of /dka ‘heal’). Second, there were clusters that can be derived from /gt/ sequences, in words such as byget ‘built’ (with a short vowel; supine of bygga ‘build’) or vdgt ‘weighed’ (with a long vowel; supine of vdga ‘weigh’). Phonetically, the two types of clusters were found to be very similar. In both types, the two stop components were released and voiceless, and the cluster tended to be preceded by a slight preaspiration. The mean ei? acini i A A total of 48 intervocalic fortis clusters were recorded, consisting of the sequences [pt] and [kt] (24 of each). These clusters were invariably produced as voiceless, with both stops released. Like simple intervocalic fortis stops, they were generally preceded by a slight preaspiration. The mean duration of this preaspiration was 31 ms. Also, mean postaspiration duration was short, 26 ms, which is not indicative of any significant postaspiration percept. An example of an intervocalic fortis cluster in képte < k6/pd/e “bought (past)’ is given in (6).
be optimal. The second tableau in (15) shows that FAITH,,.,, must be ranked above *VoI, or [k]ap would be optimal. Given richness of the Base, the grammar must map inputs with voiced spread glottis stops or with voiceless unaspirated stops in word-initial position to possible output forms. However, with only the constraints outlined so far, impossible surface forms would be designated as optimal with such inputs, as illustrated in (16). Voiced spread glottis stops are prohibited.
be optimal. The second tableau in (15) shows that FAITH,,.,, must be ranked above *VoI, or [k]ap would be optimal. Given richness of the Base, the grammar must map inputs with voiced spread glottis stops or with voiceless unaspirated stops in word-initial position to possible output forms. However, with only the constraints outlined so far, impossible surface forms would be designated as optimal with such inputs, as illustrated in (16). Voiced spread glottis stops are prohibited.
AGREE » FAITH oi In (21) we illustrate how progressive devoicing is accomplished with these ranked constraints. The first candidate is eliminated because there are no laryngeal specifications on the stops. The second is eliminated because the stops do not agree in laryngeal specifications, and the third candidate is eliminated because it violates the faithfulness constraint on [spread glottis]. A candidate with a voiced, aspirated stop would be excluded by the high-ranked constraint against voiced, spread glottis stops which we omit from the tableaux. The tableau in (21) shows that AGREE must be ranked above FAITH ,,,).-
AGREE » FAITH oi In (21) we illustrate how progressive devoicing is accomplished with these ranked constraints. The first candidate is eliminated because there are no laryngeal specifications on the stops. The second is eliminated because the stops do not agree in laryngeal specifications, and the third candidate is eliminated because it violates the faithfulness constraint on [spread glottis]. A candidate with a voiced, aspirated stop would be excluded by the high-ranked constraint against voiced, spread glottis stops which we omit from the tableaux. The tableau in (21) shows that AGREE must be ranked above FAITH ,,,).-
(22) regressive We are assuming that aspiration (whether preaspiration or post-aspiration) is the phonetic realization of the feature [spread glottis]. However, postaspiration does not occur on a [sg] stop that precedes an obstruent, and preaspiration does not occur on a [sg] stop that follows an obstruent. Hence, some segments will be specified as [sg] which are not preaspirated because of their position in the word.
(22) regressive We are assuming that aspiration (whether preaspiration or post-aspiration) is the phonetic realization of the feature [spread glottis]. However, postaspiration does not occur on a [sg] stop that precedes an obstruent, and preaspiration does not occur on a [sg] stop that follows an obstruent. Hence, some segments will be specified as [sg] which are not preaspirated because of their position in the word.
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved