Full text of DOI:10.1017/S0305000900004451

J. Child Lang.  (), –. Printed in the United Kingdom
#
 Cambridge University Press
Discriminating between constructivist and nativist
positions: fillers as evidence of generalization
(
EWA DABROWSKA
University of Sheffield
In her note, Ann Peters distinguishes between two types of fillers: () pre-
morphological, whose presence in the utterance is motivated by purely
phonological considerations (the child attempts to produce an adult utterance
at the prosodic level, or approximates the prevalent rhythmic pattern in the
language); and () protomorphological, which function as place-holders for
grammatical morphemes and eventually differentiate into various gram-
matical morphemes. I would like to suggest that there may also be a third
type which sometimes appears when the child begins to generalize over a set
of related words. Just as protomorphological fillers can provide valuable
information about how children acquire function words, these ‘generalized’
fillers may offer insights into how children form categories of function words.
I will illustrate this phenomenon with data from the Naomi corpus in the
CHILDES database (Sachs, ).
When she was ; Naomi began to produce utterances with filler syllables
consisting of a neutral vowel which was sometimes followed by an alveolar
fricative. In the note on Naomi’s lexicon accompanying the corpus, these
were described as phonologically consistent child forms, and they were
orthographically transcribed as uh and uhs. Many of the utterances containing
fillers were questions (e.g. uhs!kitty doing? uh!a passy?), but she also used
the same syllables in the preverbal position in declaratives (where the
intended meaning appears to be either ‘I’ or ‘want’). Filler syllables were
reasonably common in Naomi’s speech during the period from ; to ;
(about n% of all words), and then disappeared. The following remarks will
concentrate on the use of fillers in interrogative utterances (i.e. sentences
which end with a question mark in the transcripts).
What is interesting about these syllables is that they appeared in Naomi’s
speech at a time when she was already using the adult forms of the
morphemes that should appear in these positions, although at this stage her
use of these forms was restricted to a small number of highly formulaic and
hence presumably rote-learned phrases. Thus, they were clearly not pre-
cursors of grammatical morphemes. Furthermore, the fillers did not displace
the adult-like forms: both forms co-existed during this period (cf. examples
()–()). Thus, the appearances of these ‘empty’ syllables marked a tem-



(
porary dip in performance, in that the utterances with fillers are less adult-
like than the earlier utterances without them.
(
(
(
(
(
(
(
) (a) ;n did Daddy fall down?
b) ;n uh! Nomi fell down?
) (a) ;n is toaster?
b) ;n uh! a passy!?
) (a) ;n can I draw the cart?
b) ;n uh! draw this?
) (a) ;n what’s donkey doing?
b) ;n uhs! lamb doing?
) (a) ;n what’s Daddy got?
b) ;n uh! pinger!got?
) (a) ;n where salt go?
b) ;n uh! skunk go?
) (a) ;n where’s Sandy’s pants?
b) ;n uh! Georgie blanket?
(
(
(
(
(
(
(
Why then did Naomi use them? It seems uh and uhs functioned as a kind
of question marker. They were always utterance-initial, and, with one
exception (uh! where going?), they always appeared where the auxiliary or
the WH-word would have been in the adult version of the utterance.
Furthermore, Naomi sometimes replaced adults question markers with fillers
in imitations.
(
) ;n *FAT: what-‘is the kitty doing TNomi?
NAO: uhs! kitty doing?
) ;n *MOT: what does she want?
NAO: uhs! she want?
*
(
*
However, the fillers occurred in a variety of question types, both WH and
Y\N, and hence they were distributionally more promiscuous than any adult
form. Phonologically, they can be regarded as the ‘lowest common de-
nominator’ of all question markers, since all of the latter contain a vowel, and
many end in -s (e.g. is, does, what’s, where’s: at this stage, forms such as what’s
or where’s were unanalysed amalgrams). Thus, the fillers in Naomi’s
questions appear to be a kind of generalized question marker which captures
the features shared by all the words occurring in a particular position and
expressing a particular meaning. Her use of these ‘empty’ syllables suggests
that she was aware that interrogative utterances share certain properties – in
other words, that she had begun to generalize over her repertoire of rote-
learned phrases. Furthermore, she was aware that some kind of explicit
interrogative marker was required, but was either temporarily unable to
access it or else was not sure which of several forms that could occur in this
position was to be used on that particular occasion.



    
I suggest that the use of such ‘generalized’ fillers deserves special attention
because they are a unique source of evidence on the emergence of grammatical
categories, particularly relevant to discriminating between the nativist and
the constructivist position on this issue (positions a and b in Peters’ note).
Two observations about Naomi’s early usage are especially interesting in this
connection. First, Naomi used filler syllables in place of preposed auxiliaries
and WH-word-plus-auxiliary amalgams, but not auxiliaries in declaratives.
Thus, if her use of fillers is indeed evidence of an emergent grammatical
category, this category does not correspond to any category of adult
grammar. Secondly, as noted earlier, the vast majority of fillers first appeared
in contexts where the adult forms of the auxiliary or the WHjAux amalgram
were already well established. If she already had adult auxiliary and WH
categories and was merely learning their various phonetic realizations, we
would expect the majority of fillers to occur in sentences which require a
novel form. In fact, only two of Naomi’s interrogative fillers occur in
positions which would require an auxiliary form which is not yet well
established in her speech, and, as far as it is possible to infer her meaning
from the context, there are no instances where the emergence of a new form
was foreshadowed by a filler. Both of these facts lend support to the
constructivist view that grammatical categories are not available a priori, but
are acquired by generalizing over previously learned instances.
Of course, the data is open to other interpretations, and even if the
interpretation suggested here is correct, we cannot draw any firm conclusions
on the basis of evidence from a single child. Furthermore, the course of
development seen in Naomi, where filler syllables appeared at a time when
the child was already using adultlike forms, is unusual: I do not know of any
other child who followed a similar trajectory. On the other hand, such
children are likely to go unnoticed, precisely because the developmental
pattern is rather unexpected. There are many reports in the literature of
fillers alternating with more adultlike pronunciations. When no further
information is available, we naturally assume that the filler appeared first and
was gradually displaced by the target form – but this may not always be the
case. If they are found in other children as well, generalized fillers could
provide a unique source of evidence on the emergence of grammatical
categories, and thus they deserve special attention.
REFERENCES
Sachs, J. () Talking about the there and the then: the emergence of displaced reference
in parent-child discourse. In Keith E. Nelson (ed.), Children’s Language, Vol. . Hillsdale,
NJ: Erlbaum.
