{ "paper_id": "Y99-1024", "header": { "generated_with": "S2ORC 1.0.0", "date_generated": "2023-01-19T13:37:14.101021Z" }, "title": "SIMPLE SYLLABLE STRUCTURE AND ENGLISH PRE-NUCLEUS GLIDES", "authors": [ { "first": "Hypo-Young", "middle": [], "last": "Kim", "suffix": "", "affiliation": { "laboratory": "", "institution": "University of Michigan", "location": { "settlement": "Ann Arbor", "country": "U.S.A" } }, "email": "hyk@umich.edu" } ], "year": "", "venue": null, "identifiers": {}, "abstract": "The purpose of this paper is to test whether the Simple Syllable Structure (SSS) proposed by Duanmu (1990:[2]) is true of English onsets or not. If the SSS holds for English onset, English onglides should be secondary articulations on the onset consonant rather than separate sounds. We will test this by measuring the duration of the initial consonant-glide combination, to determine whether it is as long as two sounds or as one sound. If the consonant-glide combination is as short as one consonant, we can say that glides are secondary articulations supporting a simpler syllable structure hypothesis (Duanmu 1990:[2], Kim 1997:[6]) for English onsets.", "pdf_parse": { "paper_id": "Y99-1024", "_pdf_hash": "", "abstract": [ { "text": "The purpose of this paper is to test whether the Simple Syllable Structure (SSS) proposed by Duanmu (1990:[2]) is true of English onsets or not. If the SSS holds for English onset, English onglides should be secondary articulations on the onset consonant rather than separate sounds. We will test this by measuring the duration of the initial consonant-glide combination, to determine whether it is as long as two sounds or as one sound. If the consonant-glide combination is as short as one consonant, we can say that glides are secondary articulations supporting a simpler syllable structure hypothesis (Duanmu 1990:[2], Kim 1997:[6]) for English onsets.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Abstract", "sec_num": null } ], "body_text": [ { "text": "1. THORETICAL BACKGROUND", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "The Simple Syllable Structure (SSS) was proposed for Chinese by Duanmu (1990: [2] ) and adopted for Korean by Kim (1997: [6] ). As shown in (1) , in the SSS maximally three positions are allowed: one for the onset, one for the nucleus, and one for the coda. As the first step to test whether this SSS is true of English onsets, I measured the consonant-glide combinations (CG) in the onsets. To see why CG's are chosen, we need to look at the assumption by Feature Geometry on which the SSS depends.", "cite_spans": [ { "start": 64, "end": 81, "text": "Duanmu (1990: [2]", "ref_id": null }, { "start": 110, "end": 124, "text": "Kim (1997: [6]", "ref_id": null }, { "start": 140, "end": 143, "text": "(1)", "ref_id": "BIBREF0" } ], "ref_spans": [], "eq_spans": [], "section": "Simple Syllable Structure", "sec_num": "1.1" }, { "text": "According to Feature Geometry (FG), each articulator can only occur once in a single sound and thus non-homorganic clusters can be represented as one sound while homorganic ones cannot. For example, [t] and [w] can be represented with one feature tree because they are articulated by two different articulators, namely, Coronal and Labial while [t] and [j] cannot because both of them are articulated by a same articulator Coronal. However, it should be noted that FG does not say the non-homorganic sounds have to form one sound but says that they can form one sound. Feature trees for homorganic and non-homorganic clusters are given in (2). being the secondary articulation is that a secondary articulation is more vowel-like and a primary articulation is more consonant-like t .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Feature Geometry", "sec_num": "1.2" }, { "text": "About the duration of a secondary articulation, Ladefoged and Maddieson (1996: [4] ) note that the total duration of the gestures in secondary articulations does not equal that of a sequence of two corresponding articulations in a cluster. The proposed reason for this is that in the case of secondary articulations, the two gestures occur largely simultaneously, even though there is a small timing offset between the articulations. However, I am not aware of any study of durational measuring of consonant clusters in English except Klatt (1978) . However, his study does not include consonant + the glide /j/ and does not compare the duration of consonant clusters (CC) and single complex consonants with a secondary articulation ( C c ).", "cite_spans": [ { "start": 48, "end": 82, "text": "Ladefoged and Maddieson (1996: [4]", "ref_id": null }, { "start": 535, "end": 547, "text": "Klatt (1978)", "ref_id": "BIBREF5" } ], "ref_spans": [], "eq_spans": [], "section": "Issue", "sec_num": "1.3" }, { "text": "Most English onset clusters are non-homorganic CC's as is seen in (3). In (3), `+' indicates a possible combination and `-' indicates an impossible combination.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "ENGLISH ONSET", "sec_num": "2." }, { "text": "(3) American English Onsets: two consonants (Kenstwicz 1994:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "ENGLISH ONSET", "sec_num": "2." }, { "text": "[5])2 CI\\ C2 w j r 1 m n p t k p- + + + - - - - - t + - + - - - - - - k + + + + - - - - - b- + + + - - - - _ d + - + - - - - _ _", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "ENGLISH ONSET", "sec_num": "2." }, { "text": "For more details, see Ladefoged and Maddieson (1996: [4] ). In British English [tj] is permitted as in tune.", "cite_spans": [ { "start": 22, "end": 56, "text": "Ladefoged and Maddieson (1996: [4]", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "ENGLISH ONSET", "sec_num": "2." }, { "text": ". . . 3. EXPERIMENT", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "ENGLISH ONSET", "sec_num": "2." }, { "text": "g + + + + - - - - _ f- + + + - - - - _ e - + - - - - - - S- - + . _", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "ENGLISH ONSET", "sec_num": "2." }, { "text": "To see if a pre-nucleus glide is a secondary articulation on the primary consonant of the onset or not, a small-scaled acoustic analysis was performed. For the experiment I chose CG's among other onset CC's because all CG's are non-homorganic and thus can be represented with one feature tree. This study would be the first step to test whether the SSS is true of English onset or not, at least of CG's. The experiment contains a set of comparisons. In each set we measure the duration of the onset of one syllable3 words where the onset is a consonant + glide, a consonant alone, or a glide alone: CG in # CGV(C), C in # CV( C ), and G in # GV( C ). To take an example, the durations of /tw/ in a word twin, /t/ in tin, and /w/ in win are measured. If the glide is a secondary articulation, the duration of the CG's should be shorter than the sum of the duration of C and G when they appear separately, approximating the duration of one consonant. The duration of CG is There are two disyllabic words in the word list because I could not find appropriate monosyllabic words. They are dwindle, and beauty. expected to be closer to that of the longer consonant of the two consonants involved or a half of the durational sum of the two consonants. In (6) a capital letter D stands for duration. ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Design", "sec_num": "3.1" }, { "text": "The list was randomized and recorded in a sound proof room at a moderate rate with five repetitions by two male speakers. All words were spoken in the carrier sentence \"Say X again\" in order to produce speaking rates closer to conversational speech and to avoid effects of pre-pausal lengthening in utterance-final position. I included both aspiration and closure into the duration of stops, which is shown spectrograms in (7). Ideally the average value of column 4 should be half of that of column 5 if C G is really one sound. In other words, we expected the average duration of C G to be close to 97.9 or 136.7 but it is 165.6. This suggests that the average duration of the consonant -glide combination is much longer than that of a single consonant (97.9 or 136.7) although it is shorter than that of two consonants in succession (195.7).", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "METHOD", "sec_num": "3.2" }, { "text": "(9) Expected Length of CG CG = 1/2 (C +G) or a longer C = 97.9 or 136.7 (ms.)", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "METHOD", "sec_num": "3.2" }, { "text": "One interesting thing is that the glide /j/ behaves differently from the glide /w/. The different behavior is clear in tables 2 below. The duration of Cj is closer to two segments while Cw is not that close. The consonant-/j/ combination is as long as the sum duration of two consonants, but the consonant-/w/ is shorter than the sum duration of two consonants.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "METHOD", "sec_num": "3.2" }, { "text": "This different behavior of /j/ and /w/ is not supportive of Feature Geometry and the Simple Syllable Structure hypothesis. It is of interest to note that Davis and Hammond (1995: [1] ) argue that /w/ in a CwV sequence is in the onset while /j/ in a CjV sequence is in the nucleus based on some phonological phenomena such as phonotactics, Pig Latin, a name game, and [j] insertion4. Interestingly enough, our finding is supportive of their claim. In addition, Duanmu (1997: [3] ) claims that the English diphthong /ju/ is a long diphthong rather than short one. This finding is also very supportive of Duanmu's claim.", "cite_spans": [ { "start": 154, "end": 182, "text": "Davis and Hammond (1995: [1]", "ref_id": null }, { "start": 460, "end": 477, "text": "Duanmu (1997: [3]", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "METHOD", "sec_num": "3.2" }, { "text": "Finally, I converted the absolute value to the relative value, setting the sum duration of two consonants as 2. This is to see,how the length of CI and C\" compared with C+ j and C + w. The converted values are given in table 3.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "METHOD", "sec_num": "3.2" }, { "text": "Duration of C + C C\"", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Converted duration", "sec_num": null }, { "text": "1.48 (<-151.4) 2 (<-204.3) 0", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Converted duration", "sec_num": null }, { "text": "1.92 (<-179.8) 2 (<-187.1) The values in table 3 are visualized in the following graph. The point 2 indicates the sum duration of two independent consonants. 4 At the Mid-Continental Workshop on Phonology, Davis S. pointed out that in his theory t and w in twin have two separate positions in the onset. He also pointed out the fact that nw is not allowed in English even though they do not share the same articulator.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Converted duration", "sec_num": null }, { "text": "tin, win, twin din, win, dwindle kick, wick, quick poor, your, pure, booty, you, beauty coo, you, cue", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "WORD LIST", "sec_num": null } ], "back_matter": [ { "text": "Judging from the analysis above, first, the duration of C G is a little bit shorter than the sum of two consonants in isolation but much longer than one consonant. Second, C'' and behave differently: the former is much shorter than two consonants and the latter is as long as two consonants. This finding does not seem to support Feature Geometry and the SSS hypothesis. The reason is that Feature Geometry and the SSS expect C w and C-1 to behave in the same way, and under the SSS both Cy' and are predicted to be as long as one segment.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "annex", "sec_num": null }, { "text": "The experiment is a small scale pilot study. There are three things, in particular, I would like to explore for further study. First, we need study including more tokens, more speakers, and statistical analysis. Second, the duration of whole words and vowels would be measured. If a glide is really an independent sound, then the duration of the whole word with a glide would be longer than words with one onset consonant, and vowel length of the words with CG and the words with one C should be the same. (10b) is the case when we can say a glide is an independent sound rather than a secondary articulation. If the case turns out to be (10c), then we need more careful interpretation about the result.For example, the duration of win or tin should three-fourths as long as twin if /w/ is not a secondary articulation. If the whole word with a glide is as long as the word without a glide, then we would not say that a glide is a separate sound, even though the duration of a combination of a consonant and a glide is longer than one sound. Third, further study would involve homorganic onset CC's, coda CC's, and non-homorganic coda CC's in English. In addition, a comparative study of the durational difference between and Cw in Korean would be of interest because in Korean both /j/ and /w/ are argued to be in the same position of the syllable structure unlike English /j/ and /w/.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "FURTHER STUDY", "sec_num": "6." } ], "bib_entries": { "BIBREF0": { "ref_id": "b0", "title": "On the status of onglide in American English", "authors": [ { "first": "S", "middle": [], "last": "Davis", "suffix": "" }, { "first": "M", "middle": [], "last": "Hammond", "suffix": "" } ], "year": 1995, "venue": "Phonology", "volume": "12", "issue": "", "pages": "159--182", "other_ids": {}, "num": null, "urls": [], "raw_text": "Davis S. and M. Hammond.1995. On the status of onglide in American English, Phonology 12:159-182. Cambridge: Cambridge University Press.", "links": null }, "BIBREF1": { "ref_id": "b1", "title": "Formal study of syllable, tone, stress and domain in Chinese languages", "authors": [ { "first": "San", "middle": [], "last": "Duanmu", "suffix": "" } ], "year": 1990, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Duanmu, San. 1990. Formal study of syllable, tone, stress and domain in Chinese languages, Ph.D. dissertation, MIT.", "links": null }, "BIBREF2": { "ref_id": "b2", "title": "The Sounds of the World's Language", "authors": [ { "first": "P", "middle": [], "last": "Ladefoged", "suffix": "" }, { "first": "Ian", "middle": [], "last": "Maddieson", "suffix": "" } ], "year": 1996, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Ladefoged, P. and Ian Maddieson. 1996. The Sounds of the World's Language. Massachusetts, MA: Blackwell.", "links": null }, "BIBREF3": { "ref_id": "b3", "title": "Phonology in Generative Grammar", "authors": [ { "first": "Michael", "middle": [], "last": "Kenstowicz", "suffix": "" } ], "year": 1994, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Kenstowicz, Michael. 1994. Phonology in Generative Grammar. Massachusetts, MA: Blackwell", "links": null }, "BIBREF4": { "ref_id": "b4", "title": "Prenucleus Glides in Korean. Presented at the 3 rd McWop", "authors": [ { "first": "Hyo-Young", "middle": [], "last": "Kim", "suffix": "" } ], "year": 1997, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Kim, Hyo-Young. 1997. Prenucleus Glides in Korean. Presented at the 3 rd McWop, Indiana University, Broomington.", "links": null }, "BIBREF5": { "ref_id": "b5", "title": "Durational characteristics of prestressed word-initial consonant clusters in English", "authors": [ { "first": "D", "middle": [ "H" ], "last": "Klatt", "suffix": "" } ], "year": 1978, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Klatt, D.H. 1978. Durational characteristics of prestressed word-initial consonant clusters in English. Quaterly Progress Report. Research Laboratory of Electronics, MIT. Cambridge, MA: The Laboratory.", "links": null } }, "ref_entries": { "FIGREF2": { "uris": null, "text": "Hypothesis a. if D(tw) = (D(t) + D(w)) or longer C of the two C's , then tw is one sound, supporting the SSS b. if D(tw) = D(t) + D(w), then tw is two sounds, not supporting the SSS The word list is given in Appendix 1. A list of the words was constructed to include 18 words.", "type_str": "figure", "num": null }, "TABREF1": { "html": null, "num": null, "content": "
,
___-_
s+--++++++
All CG's and most onset CC's in English have two different articulators and thus can have a
representation of (2a). The exceptions to (2a) are in (4). CC's in (4) cannot be represented by one
feature tree but should have two trees as in (2b).
(4) homorganic CC
alveolar + /r/ : er, Sr
/s/ + C : sl, sn
(5 )x
/\\
/\\ [-continuous]
OralPhargyngeal
/\\/
[-ant] [-dist] [-voice] [-spread glottis]
", "text": "Even though /tr, dr/ are articulated by the same articulator, Coronal, they are fine to form one sound because these are considered as retroflexed affricates in Feature Geometry. Feature geometric representation for the retroflex sound /tr/ is in(5). To get retroflex sound the only thing to be added is [-ant] to the representation of W.", "type_str": "table" }, "TABREF2": { "html": null, "num": null, "content": "
(8) a. Spectrogram of [j]
b. Spectrogram of [w]A1:0.3643 A2:0.432
WOO
3600 A
3200
b. Spectrogram of [k]4. RESULT
The average durations are given in table 1.
D(C)D(G)D(CG)D(C) + D(G)
t, w, tw146.367.4172.1213.7
d, w, dw117.367.4129.8184.7
k, w, kw149.764.8152.2_214.5
P, i, Pi b, j, bj149.7 99.849.9 52.3190.8 147.2199.6 152.1
k, j, kj157.252.3201.3209.5
average136.759.0165.6195.7
TABLE 1 : duration of CG's (ms.)
4 100 140 260 2 0 360 340 460 440 560 540 6a0 640 700 mEAQU'\u00b0
-226-
", "text": "Onset and offset point of glides are defined in combination of the spectrogram and waveform displays. As we can see in (8), glides has overall low amplitude and weak formant structure. I excluded the transition period of glides.", "type_str": "table" }, "TABREF4": { "html": null, "num": null, "content": "", "text": "Converted Length", "type_str": "table" } } } }