Rich Phonology:
Some Background Material
Robert F. Port
Indiana University

1.  Formal linguistics assumes a low-dimensional fixed phonetic segment inventory.

Chomsky, N., & Halle, M. (1968). The Sound Pattern of English.New York: Harper and Row.  Chapter 1 (3-14) and the beginning of Chapter 7 (pp. 293-309).  

Haugeland's goal was to clarify the assumptions of formalist models of cognition.  These basic assumptions are shared across computing, logic, artificial intelligence and formal models of language and cognition. Major proponents of this view include Chomsky, Newell and Simon, Fodor, Pylyshyn, Pinker, etc.  Computers are carefully engineered so that they can be interpreted in terms of discrete states only.  Programmers (unlike computer engineers) do not need to concern themselves with continuous time or continuous values of voltage. Formal linguistics always assumes an apriori set of letter-like phonetic tokens. Thus it must presume the physical existence of a fixed alphabet of phonetic symbols for spelling linguistic items, as well as a discrete computational apparatus for token manipulation.  All formal linguistic theories build upon this assumption even though it is surely very implausible and the empirical claim of a universal phonetic alphabet has scarcely even been defended by phonologists. (I am sure there must have been attempts, but I'm not familiar with them.)  Chomsky and Halle (1968) took it to be intuitively obvious that such a fixed and small inventory (under a few hundred) of universal phonetic tokens exist.  So the problem, I claim, is that no such low-dimensional description of phonetics is possible (See Port and Leary, 2005).  Therefore formal linguistics is impossible.

2.  Segments are not compatible with spectrographic representations.

3.  Variation is endless when you look close.

Labov, William. (1964) Phonological correlates of social stratification. American Anthropologist 66, 164-176.

The first paper is Labov's fascinating MA thesis and the second reviews the results of his dissertation on language variation on the Lower East Side of NYC.  Labov revealed vividly the subtlety in dialect variation.  Of course, each individual speaks a different dialect, but in fact, we each speak in many different styles of pronunciation detail.  Our theories of language presume there is a linguistic state, something we could call `The English Language' or maybe `Lower East Side English' -- in any case, some invariant Grammar.   But Labov's data show us (in my view, not necessarily his) that no such entities are definable in fact.  Everything is always subject to subtle phonetic variation.   All we can know about a language is distributional patterns in a high-D auditory/phonetic space. Traditional theories of language presume that speaker and hearer share some common Grammar. But the fact is they never ever share the same grammar!   And every word has an unlimited number of variants!    So why do linguists imagine they could describe some dialect or other in discrete low-D terms?  

 Bybee, Joan (2002)   Word frequency and context of use in the lexical diffusion of phonetically conditioned sound change.  Language Variation and Change 14, 261-290.

   Bybee and other variationists have explored the ways that the distribution of variants for any word is influenced by such properties as the frequency of occurrence of each individual word -- as well as many other factors.  If every word has a bazillion variants, what can it mean to say ``I know Word X''?  Apparently it must mean either (A) ``I know the list of (essentially) all variants'' or, perhaps, (B) ``I am able to evaluate utterances for their similarity to the whole distribution of variants of Word X.''  This does not resemble what generative phonologists are doing -- but they too think they are explicating what it means to know Word X. 

4.  Long-term memory, including memory for words, is far richer than we thought.

Hintzman, Douglas (1986)  ``Schema abstraction'' in a multiple-trace memory model. Psychological Review 93, 411-428.   

This, the granddaddy of mathematical memory models, is called Minerva2, and is easier to understand than most since the math is simple linear algebra.   You will see how a rich and detailed memory of concrete exemplars with no abstractions can perform better on training tokens yet still allow abstractions (and prototypes) to be computed on the fly whenever needed.  Many more recent models (e.g., by Nosofsky and Shiffrin) use different mathematics but retain a key role for the training exemplars to predict results. (The pic: Doug Hintzman.)

O'Reilly, Randall and Kenneth Norman (2005)  Hippocampal and neocortical contributions to memory: Advances in the complementary learning systems framework.  Trends in Cognitive Sciences 6, 505-510.  

This proposes a neural model for `episodic memory,' that is, memory of randomly coocurring events (eg, the words  you hear someone say to you, or the specific events that happened on the way to work this morning) that represents stored information on a single trial.

Gluck, M., M. Meeter and C.E. Meyers  (2003) Computational models of the hippocampal region: linking incremental learning and episodic memory.  Trends in Cognitive Science 7,  269-276.

The point of these papers is that there are neurologically plausible mechanisms that can learn details of specific examples of complex, multimodal episodes on a single trial without repetitive training.  These papers suggest that a surprising amount of rote material could be available in the memory of speaker-hearers.  Remembering all examples forever is not plausible, but we apparently remember much richer detail -- including language detail -- than most of us imagined.

              Any implications for the study of language?  Of course.

Pisoni, D. B. (1997). Some thoughts on `normalization' in speech perception. In K. Johnson & J. Mullennix (Eds.), Talker variability in speech processing (pp. 9-32). San Diego: Academic Press.   

5.  What modality are phonetic targets?:  articulatory, auditory, somatosensory?

 Liberman, Alvin M., Franklin S. Cooper, Donald P. Shankweiler and Michael Studdert-Kennedy (1968)  Perception of  speech code.  Psychological Review 74, 431-461.    Reproduced in J. Miller, R. Kent and B. Atal (eds.) (1991) Papers in Speech Communication: Speech Perception.  (Acoustical Society of America, Woodbury, New York).

A monumental review paper summarizing the Haskins Labs view of speech perception arguing for the `motor theory of speech perception'.   This group thought that words must be specified in memory in low-dimensional, segmental and articulatory terms, so the  di/du problem (illustrated at right) was major.   But if auditory memory is very large, then speakers can just store context-sensitive formant trajectories and burst spectrum shapes.  The fact that  di and du begin with what we call ``the same sound'' may be something that is noticed by most speakers only after learning to write with an alphabet.   At least this issue about what speech sounds ``sound the same'' needs more research.

   Lindblom, Björn (2004) The organization of speech movements: Specification of units and modes of control.  Paper presented at conference `From Sound to Sense' at MIT.  A survey of speech production theories. 

Speech sounds must have both convenient articulatory properties and practical auditory properties. Diehl, Lotto and Holt argue for a speech perception process using only general auditory capabilities (rather than unique specialized perceptual skills that are used for speech and nothing else).   Clearly, my position is closer to theirs than to the traditional Haskins Labs view.   Guenther's feedforward production system just produces gestures -- normally with no feedback.  But the primary realtime guidance for them is based on somatosensory (ie, orosensory) feedback (from muscles, joints and skin surfaces).   So the best answer to a question about realtime targets is: the targets are somatosensory, and neither gestural nor auditory.   So I don't agree (with Galantucci-Fowler-Turvey) that  ``perceiving speech is perceiving gestures'' (their claim number 2), it is just perceiving speech.  But their claim 3 that ``the motor system is recruited for perceiving speech'' seems to be true and provides a coherent, reliable coding method for storing speech patterns.  (See Stephen M. Wilson, 2004, Listening to speech activates motor areas involved in speech production. Nature Neurosci.)

In this paper Baddeley extends his earlier model in a useful way by added an `episodic buffer' --  a store for information from many modalities that is retained for awhile.
 

7.  Why are letters so convincing despite all the evidence?  Because of our education.

Olson, David R. (1993)  How writing represents speech.  Language and Communication 13. 1-17.   This article is  a revised version of Chapter 4 in his outstanding and highly recommended book: The World on Paper: The Conceptual and Cognitive Implications of Writing and Reading. (Cambridge:  1994).   He points out that various human notions of what language is are invariably based on whatever we represent with our writing system.  So the idea that our alphabetical writing is a `cipher' (one-to-one replacement) from phonemes into graphemes has reversed the true order!  We linguists tend to believe our language is made from phonemes primarily because we use letters to represent speech.

  Faber, Alice  (1993)  Phonemic segmentation as epiphenomenon: Evidence from the history of alphabetic writing.  In Downing, Lima and Noonan (eds) The Linguistics of  Literacy.  (Benjamins, Amsterdam) pp. 111-134.  

Faber anticipated the argument I am making here over a decade ago.  Her paper has been largely ignored, as far as I can tell.   She argued that languages seem to be composed from letter-like segments primarily because we have all been trained in alphabetic literacy.

Abercrombie, David (1949)   What is a `letter'?   Lingua 2 , reprinted in D. Abercrombie  (1965) Studies in Phonetics and Linguistics. (Oxford Univ Press, Oxford), pp. 76-85    A quick, historical look at confusion from the Greeks to modern times about whether letters (Lat. litera) are graphic tokens or `speech sounds.'   Abercrombie thought the story was very clear in 1949 -- since, of course, he assumed that phonemes were the correct cognitive units. But  it seems to me that he was still very confused himself about these issues.  Neither he nor most other phoneticians were ever able to sort out the source of our vivid intuitions about the segmental structure of speech.

These papers show quite clearly that a segmental description of speech (in terms, that is, of consonants and vowels) may be intuitively persuasive for us. But it is not universal -- indeed most humans, through our history, have not had such intuitions. It appears that some amount of alphabet training is necessary for people to hear speech in terms of Cs and Vs.   

          Ziegler, Johannes and Usha Goswami (2005) Reading acquisition, developmental dyslexia and skilled reading across languages: A psycholinguistic grain size theory.  Psych Bulletin 131, 3-39. 

This review paper strongly endorses the view that the alphabetical description of language (using phonemes or letters) is a result of learning how to do alphabetical reading and writing. It also demonstrates that the inconsistencies in the spelling of English greatly complicate the problem of learning to read.   Children taught to read Finnish or Italian can read new words as well after one year of school as English-speaking children can do after 3 years of school.  Of course, the latter's practical reading skill may still be better than a child learning to read Chinese after the same amount of training.

Studies of reading reveal clearly how an understanding of language in terms of phonemes is almost entirely a consequence of learning how to read with an alphabet.  This direction of causality agrees with my general story that segmental descriptions of speech are shaped and reinforced by extensive practice at reading and writing.  But this result is in serious conflict with the conventional views of linguists and psychologists about language.  The traditional view insists that learning a language and remembering vocabulary are skills that depend upon letter-like phonemes. Some set of letter-like units are said to be available in advance to all members of the species.  On this view, all speakers, whether literate or not, should store language using a cognitive alphabet.  But the data do not support any such predictions.

Of course, there is the question of how the alphabet could have been invented if phonemes lack this intuitive basis.  But it should be kept in mind that the early Greek alphabet was the result of roughly 3000 years of efforts in the middle east to engineer a culturally transmittable writing system that was easy to teach and use.   And the alphabet was only invented once -- as has been pointed out many times.  I would claim that those Greek `writing engineers' (standing on the shoulders of earlier middle-eastern writing engineers) actually invented the phoneme concept itself.  The phoneme is the ideal, the conceptual archetype, for which all alphabetical orthographies are just imperfect implementations.

Robert F. Port, April 9, 2008