Harvard Linguistics 110 - Class 09 Analysis Continued + Similarity

Linguistics 110 Class 9 (10/21/02)

Zhang/Öztürk/Quinn

Phonemic Analysis Continued (1) The generalization for Papago: “The palato-alveolar affricates occur before high vowels, and the alveolar stops occur elsewhere.” (2) Some points that emerge • •

Parallel behavior of phonetically similar sounds. Appearance of allophones that occur as separate phonemes in other languages (for example, English).

(3) Formalizing to achieve generality • • •

Assume underlying /t,d/: these are what you get if no rule perturbs the basic pattern. In general: the elsewhere allophone is set up as underlying form. State the rule as simply as possible, leaving out whatever is not needed. It’s good to give rules names, for easy reference. Improvise a name if you are not sure of standard terminology. Alveolar Palatalization stop  affricate  vowel       alveolar → palato-alveolar / ___ high 

(4) Notation a. C = consonant

X   b. Y = “segment having the phonetic features X, Y and Z” Z 

V = vowel c. / / ___ X / X ___

= “in the environment” = “in the environment before X” = “in the environment after X”

(5) Phonemic representations •

These show the underlying representation of the phoneme, which is what you have before rules apply.

•

They are traditionally written in slant brackets: / /

1

UR of [»bidÉZim] = /»bidim/

UR of [»ta˘pan] = /»ta˘pan/

(6) Illustrative derivations Underlying forms: Alveolar Palatalization: Surface forms:

‘split’ /»ta˘pan/ — [»ta˘pan]

‘vaccinate’ /»tˆkid/ »tSˆkid [»tSˆkid]

‘press’ /»dagßp/ — [»dagßp]

‘turn around’ /»bidim/ »bidZim [»bidZim]

(7) The “why” of alveolar palatalization • •

•

It is common for alveolars to affricate before high vowels. Examples: Japanese, Quebec French, Cockney English. High vowels have a narrow air channel, and when a /t/ is released into a high vowel, the burst is noisy (say [ti], [ta] to yourself to check). Affrication is possibly an exaggeration of this natural effect, for the purpose of rendering the /t/ more audibly distinct from “quieter” stops like /p,k/. Sometimes affrication can change the point of articulation.

(8) The “why” of allophones in general •

Ease of articulation. /aI/ Raising: /aI/ → [√I] / ___ [-voice] Nasalization: V → [+nasal] / ___ [C, +nasal] /s/ Palatalization: /s/ → [sÉS] / ___ S (optional) Chris Schaefer, Russ Schuh, miss Sheila

•

To make a phoneme more perceptually distinct from the other phonemes in a particular context. Alveolar palatalization in Papago Pre-/w/ Affrication: /t/ → [tS] / ___ w

•

in twin, twine, twice...

Helping out a neighbor. /bit/ i( — [bi(t]

/bid/ — z8 [bid8]

Underlying forms Vowel Shortening: Final Devoicing: Surface forms

V → [short] / ___ [-voice] [-son] → [-voice] / ___]word

2

(9) Rule ordering Often it is crucial to apply the rules in a particular order. Yana (American Indian, Hokan, Upper Sacramento Valley, California): Final Vowel Devoicing: V → [-voice] / ___ ]word Voicing Assimilation: C → [-voice] / ___ [-voice] /aba/ → [apa•] The Phonetic Similarity Principle (10)

Sounds that are accidentally in complementary distribution

/h/ occurs: initially: after consonants: medially before stress:

hear [hi®], Horatio [h´»®eÉISioÉU] adhere [´d»hI®] ahead [´»hEd], prohibit [p®oÉU»hIbˆt]

/N/ occurs: finally: before consonants: medially before unstressed:

sing [sIN], willing [»wIlIN] Bingley [»bINli], hunger [»h√Ng‘] singer [»sIN‘], thingy [»TINi], Singapore [»sIN´«pHç®]

(11) Ignore complementary distribution here—The Phonetic Similarity Principle [h] and [N] are never felt by speakers to be the “same sound”. Plausibly, this is because they are fantastically different phonetically. Conclusion: in phonemic analysis, we should be reluctant to group phoneticallydissimilar sounds into phonemes. (12) What is the cut-off point? Bruce Hayes (UCLA Linguistics Professor) and his 4-year old son Peter Hayes: BH: PH: BH: PH:

“Please say [kHæ/t|] backwards.” “[tHæ/k|]” “Please say [fI…] backwards.” “[…If]. ([dæ|i kˆn wi stAp duIN DIs naÉU]?)”

A couple years later: “[lIf]”

3