![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
dr_whomMerl Reagle's crossword for today provides a nifty example of how language play can provide data on some fairly abstract theoretical linguistic questions.
The theme of the crossword was "Unstressed Syllables": each of the theme entries was some common phrase with an unstressed schwa deleted. Thus the clue "Pro athlete who whipped all of his opponents?" leads to the entry CREAM ABDUL-JABBAR, and the clue "College subject for dentists?" has the answer FLOSSOPHY.
Okay. So there are two interesting cases of phonological neutralization in this puzzle. One is the entry A CUBAN SCAR (punning on CIGAR, obviously). This pun indicates that Reagle at least thinks it reasonable to identify the unaspirated [k] in words like SCAR with the phoneme /g/ rather than the phoneme /k/. Of course, English doesn't have a contrast between voiced and voiceless stops after /s/, so in principle a speaker could be free to identify stops after /s/ with either of the two available phonemes. But the more intriguing possibility is that stops after /s/ are actually underspecified—that the speaker's knowledge of the phonology of the word like SCAR doesn't contain either /g/ or /k/, but rather just an abstract entity "velar stop".
With the limited data in this crossword puzzle, we can't test that hypothesis for stop voicing neutralization after /s/ in Merl Reagle's grammar. But we can test it for another neutralization! Another theme entry in the puzzle is DRAINIUMS, punning on GERANIUMS. One thing this tells us is that Reagle is one of those speakers for whom apical stops are affricated before /r/—so the historical /d/ in a word like DRAIN has become [dʒ], or something close enough to it that it seems phonetically reasonable to say that DRAIN would be the result of contracting GERANiums. So does DRAIN actually contain the phoneme /dʒ/?
Well, in the same crossword we have DRECKTORIAL DEBUT, punning on DIRECTORIAL. So in Reagle's phonology, it seems as if /dʒər/ and /dər/ both produce a DR- onset when the schwa is deleted. (The probability that Reagle pronounces DRAIN and DRECK with different onsets, the former with [dʒ] and the latter with [d], I feel is sufficiently remote to be ignorable.) What, then, is the phonological status of this DR- onset? It doesn't seem like we can say anything like "In Reagle's grammar, /dʒ/ can appear in an onset before /r/ but /d/ can't," or vice versa, the way we might want to say that /t/ can't appear in an onset before /l/. If we could, then that would mean that either DIRECTORIAL or GERANIUMS ought to be uncontractable, the way TELEMETRY seems to be uncontractable (you can't say TLEMETRY). But neither one is uncontractable: they just both produce the same DR- onset when contracted. This seems to be fairly strong evidence that the first consonant in DRAIN and DRECK in Reagle's grammar is actually synchronically underspecified between /d/ and /dʒ/—neither one nor the other, but equally both of them.
The theme of the crossword was "Unstressed Syllables": each of the theme entries was some common phrase with an unstressed schwa deleted. Thus the clue "Pro athlete who whipped all of his opponents?" leads to the entry CREAM ABDUL-JABBAR, and the clue "College subject for dentists?" has the answer FLOSSOPHY.
Okay. So there are two interesting cases of phonological neutralization in this puzzle. One is the entry A CUBAN SCAR (punning on CIGAR, obviously). This pun indicates that Reagle at least thinks it reasonable to identify the unaspirated [k] in words like SCAR with the phoneme /g/ rather than the phoneme /k/. Of course, English doesn't have a contrast between voiced and voiceless stops after /s/, so in principle a speaker could be free to identify stops after /s/ with either of the two available phonemes. But the more intriguing possibility is that stops after /s/ are actually underspecified—that the speaker's knowledge of the phonology of the word like SCAR doesn't contain either /g/ or /k/, but rather just an abstract entity "velar stop".
With the limited data in this crossword puzzle, we can't test that hypothesis for stop voicing neutralization after /s/ in Merl Reagle's grammar. But we can test it for another neutralization! Another theme entry in the puzzle is DRAINIUMS, punning on GERANIUMS. One thing this tells us is that Reagle is one of those speakers for whom apical stops are affricated before /r/—so the historical /d/ in a word like DRAIN has become [dʒ], or something close enough to it that it seems phonetically reasonable to say that DRAIN would be the result of contracting GERANiums. So does DRAIN actually contain the phoneme /dʒ/?
Well, in the same crossword we have DRECKTORIAL DEBUT, punning on DIRECTORIAL. So in Reagle's phonology, it seems as if /dʒər/ and /dər/ both produce a DR- onset when the schwa is deleted. (The probability that Reagle pronounces DRAIN and DRECK with different onsets, the former with [dʒ] and the latter with [d], I feel is sufficiently remote to be ignorable.) What, then, is the phonological status of this DR- onset? It doesn't seem like we can say anything like "In Reagle's grammar, /dʒ/ can appear in an onset before /r/ but /d/ can't," or vice versa, the way we might want to say that /t/ can't appear in an onset before /l/. If we could, then that would mean that either DIRECTORIAL or GERANIUMS ought to be uncontractable, the way TELEMETRY seems to be uncontractable (you can't say TLEMETRY). But neither one is uncontractable: they just both produce the same DR- onset when contracted. This seems to be fairly strong evidence that the first consonant in DRAIN and DRECK in Reagle's grammar is actually synchronically underspecified between /d/ and /dʒ/—neither one nor the other, but equally both of them.
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
no subject
Date: 2012-06-26 05:24 pm (UTC)no subject
Date: 2012-12-13 08:49 pm (UTC)no subject
Date: 2012-12-13 08:54 pm (UTC)