Back in October I noted a number of sessions that were scheduled at the LSA last week and that might be of interest to phonologists. I’d like to invite anyone who attended any of these sessions (or anything else of phono-interest at the LSA) to offer their thoughts here on phonoloblog.
The teaching term began for me as soon as I got back from LSA, so I haven’t yet had time to write up some thoughts I had on the ominously-titled plenary panel “Phonology: An Appraisal of the Field in 2007”, but I will definitely get around to it soon.
I was at the preview session for the appraisal of phonology. Bruce Hayes kept talking about “prosthetic children” as some kind of thought experiment. Does anyone know what a prosthetic child is?
Also, one guy (whose name I’d probably know if I were a phonologist) mentioned how the phoneme was “unequivocally wrong,” to the audience’s silent assent. Gosh, I thought, I know that knowledge is constantly evolving, and some concepts go by the wayside, but I had no idea that the idea of a phoneme was not just in need of some tweaking, not just in need of serious revision or rethinking, but completely, unequivocally wrong. And no one on the panel even batted an eye. I stood up and asked why the idea of a phoneme was so wrong, but no one provided me a coherent answer. A guy that I later learned was Ian Maddieson said you didn’t need to look very far at all before problems with the concept of phoneme started cropping up, but didn’t elaborate. Could someone enlighten me? And if phonemes are out, how do you refer to sets of phones that speakers of a given language perceive to be the same sound?
Neal, I wish I could help you … all I can do at the moment is refer you to various points of view expressed here in a series of comments about the role or place of the phoneme in theory and pedagogy.
There were three sessions I attended and remembered something about: Approaches to Language Complexity, Laryngeal Features, and Sound Change.
In the session on Language Complexity, the focus was an examination of the perspective that “all languages have equal complexity.” This idea crops up often in the literature, but is usually taken to mean: “all languages have equal value,” since we as linguists often value complexity in a linguistic system. However, an investigation into this question of correlated complexity was examined by the panel.
Ian Maddieson talked about how complexity is correlated across languages, with particular reference to tonal and consonantal inventories. In sum, languages with many tonal contrasts have larger consonant inventories, while languages with fewer tonal contrasts have smaller consonant inventories. It therefore shown that tone and consonant inventory are not inversely proportional, contrary to what many might think.
In the laryngeal feature session (where I gave my talk), two of the talks focused on the “fortis-lenis” distinction in different languages (Korean and Trique), while the last talk (Avelino & Shin) discussed the realization of glottalization in Yucatec Maya. In Trique, the main correlate of the fortis consonants is duration, but for the stops there are other pertinent cues, such as the degree of glottal spreading during closure and increased F0 upon release (in a tone language with 9 tones!).
In the session on Sound Change, Alan Yu presented an interesting talk on how speakers favor sound change moving in certain directions over others. I don’t remember too many of the details though.
Unfortunately, I didn’t get to see too many talks, but from what I saw and who I talked to, it seemed to me that there was a substantial emphasis on the integration of phonetic investigation with phonological argumentation. This was mentioned during the Phonology: Appraisal of the Field session too.
Neal — I forget who it was, but someone in the audience clarified that Bruce must have meant synthetic children. (Don’t know whether that clarifies to you what he was talking about, though.)
The issue of the phoneme being “wrong” is complicated, in part (I think) because there are two very different things going on when people talk about it.
First, for many it’s a reference to Halle’s early argument that the structuralist phonemic level (midway between the morphophonemic and phonetic levels, the relation between each pair respecting biuniqueness) gets in the way of expressing linguistically significant generalizations. (The importance of this argument is mentioned around the middle of this Language Log post; there’s a link at the end to this summary of Halle’s argument). The result is that there is no phonemic level, therefore no phoneme, but subsequent generations of phonologists have simply adopted the term “phoneme” to refer to the elements of underlying representations, which are nevertheless closer to the structuralist morphophonemic level.
Second, I think that at least some people may be thinking of the issues raised in some of the comments that Bob has already pointed to. In particular, several strands of work in OT (e.g., exploring the consequences of the Richness of the Base hypothesis) have focused attention on the content and utility (or if you prefer, the nature/status/rôle) of underlying representations — which, as I’ve just noted, phonologists tend to think of as being composed of phonemes. I also plan to say more about this here soon, especially with regard to Larry Hyman’s talk at the appraisal-of-phonology session at the LSA.
There is perhaps another view on the “invalidity” of the phoneme not discussed here. A view that is common in the psychological literature, where many practitioners want to posit as little abstraction as possible (sort of a revival of the old associationist view of cognition). In many cases, models are proposed with “no” abstraction (i.e., the types of exemplar/episodic models espoused by people like Stephen Goldinger (ASU) and some of the work by the late Dennis Klatt, among many, many others). For them, lexical representations are simply spectrograms or some acoustic representation that is very transparent to the raw input signal. There is also work summarized in an old paper by Jacques Mehler, with John Morton and Peter Jusczyk (“On Reducing Language to Biology”, Cognitive Psychology 1(1): 83-116), where they talk about old experiments (some done by Alvin Liberman) that were unable to find psychological evidence for the phoneme. Take what you will regarding conclusions drawn from such experiments, but there is a strand of thought in psychology that rejects the notion of the phoneme and other abstract representations.
While I find these ideas (rejection of abstraction) mystifying, given the good linguistic (and newer psycholinguistic) evidence we have (which is also underappreciated outside of linguistics), they do represent the position of many, many psychologists and neurobiologists.
While I certainly don’t advocate such a position, you can read some more about it in:
Goldinger, Stephen D. and Tamiko Azuma (2003). “Puzzle Solving Science: The Quixotic Quest for Units in Speech Perception.” Journal of Phonetics 31: 305-320.
Pingback: phonoloblog»Blog Archive » Erculator