Shadow

Flaps is notoriously hard in spelling. Becky Treiman and others have shown that young children (at some point during their literacy development) tend to spell as they hear the sounds, such as <*SDAR> for <STAR>. Does Eddington suggest here that children actually have an "underlying" lexical representation of /t/ in STAR? How would that come about — before one learns to spell?

Duke doesn’t seem to have online access to this journal, but fortunately the author posted a draft online.

Flaps and other variants of /t/ in American English: Allophonic distribution without constraints, rules, or abstractions

Flaps and other variants of /t/ in American English: Allophonic distribution without constraints, rules, or abstractions

---

Author(s): David Eddington
doi: 10.1515/COG.2007.002
View PDF article (248 K)		View PDF with links (252 K)

Abstract text

The distribution of the flap allophone [] of American English, along with the other allophones of /t/,[t^h,t⁼, , t] has been accounted for in various formal frameworks by assuming a number of different abstract mechanisms and entities. The desirability or usefulness of these formalisms is not at issue in the present paper. Instead, a computationally explicit model of categorization is used (Skousen 1989, 1992) in order to account for the distribution of the allophones of /t/ without recourse to such formalisms. The simulations that were carried out suggest that they are not needed because analogy to surface apparent variables such as phones and word boundaries is sufficient to predict allophony.

In analogy, the particular allophone of /t/ (i.e. [, t^h, t⁼, , t]) that appears in a given context is determined on the basis of similarity to stored exemplars in the mental lexicon. From an acquisitional standpoint, categorization by analogy to stored exemplars dispenses with the need for rule induction although it does suggest that speakers group functionally related sounds into mental categories, a process that is influenced to a great deal by orthography.

Analogy also explains the stochastic nature of linguistic performance. In the present study, 3,719 tokens of the allophones of the phoneme /t/ were extracted from the TIMIT corpus and constitute the database from which analogs were chosen. The variables used included the three phones or boundaries on either side of /t/, and the stress of the syllables preceding and following /t/. The model proves quite successful in predicting the correct allophone, and the errors made are generally possible alternative pronunciations (e.g. moun[]ain, moun[t^h]ain). The success rate changes little when only small sub-samples of the database are incorporated. In addition, exemplar-modeling is found to be quite robust because even when a feature such as stress is eliminated, (a feature which is critical in most rule approaches), allophony is still highly predictable.

I skimmed through the paper. I have to say I am a littel disappointed. The crux is that one can determine the allophone of /t/ in a particular context by comparing it to other contexts in your vocabulary (in this stimulation study, cases from a corpus with multiple speakers). And you can do it reasonably well — 65% overall, but it’s hard to compare with rule-based benchmarks, as the author argues.

How do you get the rest of the /t/ correctly? By comparing to the rest of the rest of the contexts? How is this different from a pure lexical account — i.e., you say "lader" for <later> because that’s how you remembered it? 

In some (hopefully not too cynical) ways any theory involving underlying representations and rules/constraints are just long-winded ways to account for surface similarities. But the hope is that these theoretical apparatus will help to clean up the messy surface data. I think this paper has not convinced me that they are unnecessary. 

And how does this analogy-based account explain orthographic influences?  

This entry was posted on Friday, July 6th, 2007 at 12:32 pm and is filed under Research, Paperville, Linguistics. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site. Edit this entry.