Journal of Memory and Language 44, 568–591 (2001)
doi:10.1006/jmla.2000.2756, available online at http://www.academicpress.com on
0749-596X/01 $35.00
Copyright © 2001 by Academic Press
All rights of reproduction in any form reserved.
568
Determinants of Wordlikeness: Phonotactics or Lexical Neighborhoods?
Todd M. Bailey
University of Oxford, Oxford, United Kingdom
and
Ulrike Hahn
University of Wales, Cardiff, United Kingdom
Wordlikeness, the extent to which a sound sequence is typical of words in a language, affects language acqui-
sition, language processing, and verbal short-term memory. Wordlikeness has generally been equated with
phonotactic knowledge of the possible or probable sequences of sounds within a language. Alternatively, word-
likeness might be derived directly from the mental lexicon, depending only on similarity to known words. This
paper tests these two cognitively different possibilities by comparing measures of phonotactic probability and lex-
ical influence, including a new model of lexical neighborhoods, in their ability to explain empirical wordlikeness
judgments. Our data show independent contributions of both phonotactic probability and the lexicon, with rela-
tively greater influence from the lexicon. The influence of a lexical neighbor is found to be an inverted-U-
shaped function of its token frequency. However, our results also indicate that current measures are limited in
their ability to account for sequence typicality. © 2001 Academic Press
Key Words: wordlikeness; phonotactics; token frequency; lexical neighborhood; sequence typicality.
Any speaker of English can tell that Zbigniew
is not an English name. It is not that Zbigniew
merely happens to not be in a mental dictionary
of English names; there is a strong intuition that
it could not be included in any such dictionary.
Even among actual words of English, some
words sound more typical than others, for exam-
ple rat as opposed to sphere, splurge, or flail.
Speakers have consistent intuitions about whether
a sequence of speech sounds could be a word of
their native language and how typical it would
sound as a word. Moreover, speech processing
is highly sensitive to the typicality of sound se-
quences, as we will review. These effects can be
seen in speech perception and production, lexi-
cal development, lexical access, and verbal
memory. However, it is far from clear what
makes a sequence of sounds more or less typical,
more or less wordlike. Our own desire to under-
stand the knowledge structures underlying word-
likeness was sparked initially by unsuccessful at-
tempts to use common measures of wordlikeness
for the principled construction of experimental
stimuli. We were seeking an objective method of
choosing sets of one-syllable words and non-
words covering a range of typicality, as would be
desirable in a wide variety of psycholinguistic
studies. Contrary to our expectations, standard
measures of typicality did not generally corre-
spond to our intuitions of wordlikeness in any
real way. This prompted us to examine more
closely the determinants of wordlikeness.
Sound sequence typicality is most often
thought of as phonotactic probability, that is, the
frequency with which a particular phoneme or
phoneme sequence occurs in a language. Phono-
Todd Bailey was supported by grants from the
McDonnell–Pew Centre for Cognitive Neuroscience, Ox-
ford, and a grant to Kim Plunkett from the Biotechnology
and Biological Sciences Research Council, UK. We thank
John Coleman, Gordon Brown, and Dan Jurafsky for valu-
able discussions, and we are grateful to our anonymous
reviewers for helpful comments. A preliminary analysis of
Experiment 1 appeared in the Proceedings of the 20th An-
nual Meeting of the Cognitive Science Society.
Address correspondence and reprint requests to Todd
Bailey or Ulrike Hahn, School of Psychology, Cardiff Uni-
versity, P.O. Box 901, Cardiff CF10 3YG, United Kingdom.
E-mail: baileytm1@cardiff.ac.uk or hahnu@cardiff.ac.uk.