SLM | The Speech Learning Model

The Speech Learning Model
Discussion Topic 1:
The Speech Learning Model (SLM)
If you are still interested in the SLM after reading all this (!) you might want to explore how the SLM
developed and was slowly modified over time as new empirical data were obtained. The first full presentation
of the SLM is found in a book edited by W. Strange in 1995. However, outlines of the model are to be found
in chapters published in H. Winitz (Ed.) 1988 and R. Kent (Ed.) 1992. The chapter in Winitz is quite long and
presents nearly all of the research available (in English, at least) up to about 1985 when it was written.
The chapter in volume 9 of the Laboratory Phonology series was a long look backwards
of nearly 25 years of research focused on L2 speech learning. Shortly after the chapter
appeared in 2007 I wrote a commentary on it, having had a couple more years to mull
over its contents. The commentary was meant to stimulate discussion. If you have
something to say, this would be a perfectly appropriate forum. Click
here to read my
commentary. I look forward to hearing from you.
If you have just entered the field of second-language (L2) speech research you
probably know all about the Critical Period Hypothesis already, but you may know little
about the Speech Learning Model. If you want to learn more about the SLM I suggest
you start by having a look at the keynote I gave at the Acoustical Society of America
(ASA) conference in Vancouver, Canada in April, 2005. It is entitled
The origins &
development of the Speech Learning Model.
The final elaboration of the SLM was presented in Laboratory Phonology 9. This
chapter, written in 2005, appeared in Cole & Hualde (Eds, 2007). I suggest that this
version be used by those wishing to evaluate hypotheses of the SLM by testing
predictions it generates.
Above: the main square
in Tuscania; the Post
The aim of the Speech Learning Model (SLM) is to account for differences in the
learnability of phonetic segments in an L2. Both production and perception are
examined. The SLM posits that the accuracy with which segments are perceived in an
L2 places an upper limit on the accuracy with which the same segments can be
produced. It posits, further, than the processes and mechanisms used in the successful
acquisition of the L1 sound system -- inluding the ability to establish phonetic
categories --remain intact across the lifespan and can be exploited in L2 speech