Linguistics 1 Homework B: Phonology Huave
Huave (pronounced [wave]) is a language isolate spoken in the state of Oaxaca, Mexico. The data below are taken from the San Mateo dialect. Before turning to the data I need to make clear three special properties of the language which are different from English. 1. The following six vowels occur in Huave words: [ i e æ ɨ o ɑ]. The vowels can be short or long — the long vowels are marked here with a following colon, e.g. [tɑndɑ:b] ‘burned’ has a short [ɑ] in the first syllable and a long [ɑ:] in the second. The [ɨ] vowel is a high central tense vowel. It resembles the sound of the second vowel in the English word roses, where the [ə] has a rather high allophone for many speakers. This sound occurs only as a stressless vowel in English. To practice making it, you can try to say [u] and then unround your lips completely and hold the vowel (try smiling as a way to make sure your lips are not rounded — smiling pulls back the sides of your mouth). Huave, along with Russian and Modern Irish (for example), has the important characteristic that most of the consonants can be either plain (regular) or palatalized. A palatalized segment has in addition to its ordinary articulatory gestures an additional movement of the tongue body upward and forward in the mouth, as if to pronounce a [ j] (y-like sound). For this reason palatalization is notated with a small superscript ʲ to the right of the affected segment. So for example in Huave [kɨtʲ] the final [tʲ] is palatalized. To English speakers a palatalized consonant sounds a bit like a consonant followed by [ j], so [mʲut] sounds like mute whereas [mut] sounds like moot. There is, however, a difference between a palatalized consonant and a consonant followed by [ j] — the proper transcription of English mute is [mjut]. The difference in articulation is that for a palatalized consonant the tongue gesture for the [ j] is simulataneous with the other constrictions of the oral cavity that occur in making the consononant. For a sequence like [mjut], the [j] follows the [m] slightly; or, put differently, the [m] ends before the [ j] does. The difference can be quite subtle, but the important point for this exercise is that although a palatalized consonant is notated with a sequence of symbols — a consonant symbol followed by [ ʲ ] — these two symbols together denote a single segment.

2.

3. 2

The segment inventory of Huave also contains the affricate [ʦ] (as in Canadian French). This sounds like a sequence of [t] and [s] but counts as a single segment, like English [tʃ] for example. In addition Huave has three other segments that involve a sequence of events during their duration: [mb nd ŋg]. These sound like the sequences of a nasal consonant and a stop, but in Huave these sequences count as a single segment — [mb nd ŋg] always behave just like an ordinary consonant like [b d g], except that they contain two phases of articulation, the first part nasal and the second part oral. Segments of this type are called prenasalized stops, and they also occur in many Bantu languages spoken in Africa. You may have encountered them already in the names of places or people, e.g. Ndjamena (capital of Chad) or Ngũgĩ wa Thiong’o (Kenyan writer), where a prenasalized stop appears at the beginning of the name. For the purposes of this problem what is crucial to understand is that the prenasalized stops, although they are notated with two symbols in sequence, are phonologically single segments. And, like the other consonants, they can appear as plain or as palatalized. The palatalized variants are notated [mbʲ, ndʲ, ŋgʲ], and they too are single segments. Finally, Huave even has a prenasalized affricate [nʦ], which sounds much like the sound(s) at the end of the English word pants. Although written with three symbols in a row, it is a single segment. And — perhaps no surprise now — there is also a palatalized prenasalized affricate [ntsʲ] (sounds a bit like the end of the English word lunch). It too is a single segment. To review: the following special segments need to be kept in mind — even though they are written with sequences of symbols, they are to be understood as single segments phonologically: [ts/ʦ, tsʲ/ʦʲ, mb, mbʲ, nd, ndʲ, ŋg, ŋgʲ, nʦ/nts, nʦʲ/nʦʲ] If you see these sequences in the data, they should always be interpreted as single segments and not as sequences. Note that we do not write [nʲdʲ], for example, putting [ʲ] after both parts of the prenasalized stop, but this is just a convention. Writing [ndʲ] as we do does not mean that one part is not palatalized and the other part is not; rather it is just to save writing more [ʲ] symbols.

Part 1. Plain and Palatalized Consonants 3 1 Determine from the data in Data Set 1 below whether plain consonants and their palatalized consonant counterparts are allophones. In other words, for example, are [p] and [pʲ ] allophones of a single segment /p/? (Note: you do not have to consider each pair individually: to make the problem easier, assume that either each such pair is allophonically related or none are). Or is palatalization as a property contrastive in Huave? Explain your reasoning.

Remember: to show that two segments x and y are not allophones of a single phoneme it suffices to show that they are not in complementary distribution. In other words, there should be cases where a two words differ only in that one has x and the other has y. For example English [bɪt] and [pɪt] are different words, and this establishes that [b] and [p] are not allophones of the same phoneme, but rather the difference between them is contrastive — it serves to distinguish distinct forms in English. To show that two segments x and y are allophones it is necessary to show that x and y are in complementary distribution: x never occurs in precisely the same environment that y does. To show this it is normally necessary to identify explicitly the environment where all instances of x occur, and to establish that y never occurs there (at least as far as you can determine from the data).

4

Data Set 1 ɑhpotʲ ɑlo:gʲ ɑlomb ɑloŋgʲ ɑmbʲol ɑnot ɑpolʲ ɑso:gʲ ɑsokʲ ɑtoŋg ɑsotʲ ɑtʲoʦ how ‘to bloom’ ‘extracts, pulls out’ ‘praies’ ‘rolls up sleeve’ ‘to help’ ‘to pull, to take up’ ‘injures’ ‘to intercalate’ ‘to be told’ ‘to empty, to tip over’ ‘sin’ ‘thinks’ ‘oh!’ kotʲ mbʲoʦ mbolʲ mbol molʲ montʲok moŋgʲ nʲepotpot ndo:gʲ okʲ tʲoŋg tot totʲ ‘fence, wall’ ‘blue corn’ ‘jack’ (type of fish) ‘fear’ ‘foreigner’ ‘lightning’ ‘uneven terrain’ ‘rough’ ‘to break, cut’ ‘cloud’ ‘toad’ ‘bone marrow’ ‘hip’

Part 2. [ɑ] vs. [o] 2. Use the data already presented in Data Set 1 along with the new data given in Data Set 2 below. Determine if [ɑ] and [o] are allophones of a single phoneme or if they are contrastive in Huave. Explain your reasoning. Data Set 2. nɑpɑt hɑw tɑt nʲelɑmb ɑpɑl ‘fierce’ ‘chin’ ‘white mangrove’ ‘conceited’ ‘opens’ nɑ:b tɑw pɑwɑl ɑk ‘drum’ ‘departed’ ‘round’ ‘right hand’

5 Part 3. [ɑ] vs. [ɨ] vs. [æ] 3. Data Set 3 Introduces some some words with [ɨ] and [æ], as well as more data with [ɑ]. Using the data from Data Set 3 as well as that in Data Set 2, answer the following questions: Consider the distribution of the vowels [ɑ ɨ æ] in Huave. Are these three vowels allophones of a single phoneme or are they contrastive vowels (not derived from a single phoneme)? Explain your reasoning. Here you have three segments — [ɑ ɨ æ] — to be concerned with. So if they are all allophones of one phoneme, then they should have completely nonoverlapping distributions — none occurs where either of the other does. If you choose the contrastive option, you must show pairs of words which are distinguished only by the difference between [ɑ ɨ], [ɨ æ] or [ɑ æ]. If you choose the allophones option, you need to explictly characterize the distributions of at least two of the three vowels in question.

Data Set 3: [ɑ ɨ æ] ɑndʲæk ɑsʲæ:ndʲ ɑsɑp ɑtʲæm ɑwɑ:g eʦʲwʲæk hɑɾɨnʦʲ hɨmʲ ikʲæl ilʲæ:gʲ inɨ:bʲ ɨtʲ kʲæʦʲ ‘speaks’ ‘loads’ ‘takes’ ‘earthquake’ ‘to dry out’ ‘monkey’ ‘thread’ ‘crocodile’ ‘container’ ‘you open’ ‘you sell’ ‘eats’ ‘brains’ kɑ:w kɨtʲ kɨtʲ kombɨlʲ lʲæmb lɑm lɨwʲ mʲæwɑn mʲitʲæ:ts mɑʦ mbʲæsʲ mbɨpʲ mɨsʲ ‘moon’ ‘fish’ ‘fish’ ‘compadre’ ‘sling’ ‘stream’ ‘jaguar’ ‘all’ ‘priest’ ‘cot’ ‘fingernail, toenail’ ‘horsefly’ ‘canoe’

6

nʲæmʲ nʲelʲæŋgʲ nʲerɑ:r nʲeʦʲæʦʲ nɑdɑm nɑkɑnts nɑpɑk nɑpɨkʲ nɑtɑŋg nɑwʲæl nɨtʲ ntsɑh ohpʲækɨwʲ okɑs olɨkʲ olɨmbʲ omʲæ:ts omal ondʲæʦ pʲæts pʲækʲ pʲæŋg pʲæt pɨlʲ

‘yes, okay’ ‘straight’ ‘hot’ ‘stupid’ ‘big’ ‘red’ ‘strong’ ‘sandal’ ‘grown up’ ‘gelatinous’ ‘sun’ ‘gum’ ‘eight’ ‘star’ ‘tooth’ ‘wing’ ‘heart’ ‘head’ ‘hair’ ‘tortilla’ ‘shoulder’ ‘mosquito’ ‘countryside’ ‘hoe’

rɨnʦʲ sʲæn sʲærɨwʲ sʲimʲæ:ʦ sɑhndɨkʲ sɑpʲændʲ sɑts sɑtsɨndʲ sɨ:gʲ tʲæmʲæ:mʲ tʲæmpo:ʦ tʲætʲ tɑbɑ:ʦ tɑndɑ:b tɨpʲ tsɑpɑh wʲæk wʲænʲ wɑkɨsʲ wɨ:gʲ wɨlʲ wɨsʲ

‘ranch’ ‘sir’ ‘skunk’ ‘my heart’ ‘I hiccup’ ‘I press’ ‘thorn’ ‘I hit’ ‘medicine’ ‘is blowing on’ ‘church’ ‘father’ ‘let’s go!’ ‘burned’ ‘wart’ ‘thick’ ‘horn’ ‘nance’ (type of tree) ‘cattle’ ‘floodplain, lowland’ ‘where?’ ‘on’