Syntactic class influences phonological priming of tip-of-the-tongue resolution
LISE ABRAMS
0
EMILY L. RODRIGUEZ
0
0
University of Florida
,
Gainesville, Florida
During tip-of-the-tongue (TOT) experiences, phonologically related words have both facilitated and impeded word retrieval. In the present experiment, we examined whether phonologically related words' syntactic class (part of speech) is responsible for these differential effects. Sixty college students saw general knowledge questions whose answers were designated target words and responded know, don't know, or TOT. Following TOT and don't know responses, the participants saw five words, one of which was a prime. The primes contained the target's first syllable and either shared or did not share the target's part of speech. Following presentation of the primes, retrieval of the target was attempted again. Different-part-of-speech primes facilitated resolution of TOT states, whereas same-part-of-speech primes had no effect, relative to phonologically unrelated words. These results support node structure theory's most-primed-wins principle and the transmission deficit model account of TOT states and detail the importance of syntactic class in the selection of words that are candidates for speech production.
-
A tip-of-the-tongue (TOT) state is a well-known
phenomenon that consists of a temporary inability to retrieve
a known word (e.g., R. Brown & McNeill, 1966). This
frustrating experience has motivated considerable
research in the past 25 years and has led to the development
of several theories regarding why TOT states occur. One of
the predominant theoretical explanations is the
transmission deficit (TD) model, which proposes that TOT states
are caused by weakened connections between a word and
its phonology, or sounds (e.g., Burke, MacKay, Worthley,
& Wade, 1991; MacKay & Burke, 1990). The TD model
therefore suggests a solution for resolving TOT states:
Strengthening the weak phonological links will increase
retrieval of the missing word. In previous research, this
prediction has been tested by inducing TOT states,
presenting words containing various phonological features
of the unretrieved word, and then asking participants to
attempt word retrieval again. These studies showed that
phonologically related words presented during TOT states
increase word retrieval (e.g., Abrams, White, & Eitel,
2003; James & Burke, 2000; White & Abrams, 2002).
However, phonologically related words can also
inhibit or delay TOT resolution. TOT states are often
ac
This article is based on an undergraduate senior thesis written by
E.L.R. Portions of this research were reported at the 16th annual
convention of the American Psychological Society in May 2004 and at the
45th Annual Meeting of the Psychonomic Society in November 2004.
The research was partially supported by the University of Florida
University Scholars Program, an undergraduate research award granted to
E.L.R. We thank Katherine White for her helpful comments on a draft
of this article. Correspondence concerning this article should be sent to
L. Abrams, Department of Psychology, University of Florida, P.O. Box
112250, Gainesville, FL 32611-2250 (e-mail: ).
companied by blockers, or persistent alternatesthat
is, incorrect words that come to mind involuntarily and
typically share phonological features with the unretrieved
word (e.g., MacKay & Burke, 1990). Research has
demonstrated that when TOT states are accompanied by an
alternate word, TOT states are less likely to be resolved
and that, even when resolved, retrieving the intended word
takes longer, relative to TOT states that occur without
alternate words in mind (Burke et al., 1991). These results
are consistent with an inhibition model of TOT states, in
which TOT states are caused by an alternate word, which
comes to mind first and suppresses retrieval of the desired
word (e.g., Jones, 1989).
The present experiment attempted to resolve these
paradoxical findings by testing a specific hypothesis unique
to node structure theory (NST; MacKay, 1987) and the
TD model. This hypothesis predicts that a words syntactic
class plays a pivotal role in determining the impact that
phonologically related words will have on resolution of
TOT states. Specifically, Burke et al. (1991) suggested
that for subjects in the TOT state, presenting a word that
is phonologically related and in a different domain
(syntactic class) from the target will facilitate resolution, whereas
a phonologically related word in the same domain as the
target will delay resolution (p. 570). To understand how
this prediction is derived, a brief overview of NST and the
TD model is given below.
NST arranges conceptual representations, or nodes,
into a hierarchical network of multilevel systems,
including semantic, syntactic, and phonological systems. Within
this theory, word retrieval occurs as a function of two
processes: node priming and node activation. Node priming is
a subthreshold excitation that spreads between connected
nodes, whereas node activation is the point (...truncated)