reverse mutual exclusivity?

[Gedeon Deák]

Nice example, which might say more about knowledge of language use 
rather than so-called mutual exclusivity (M.E.). I can't address the 
bilingual metalinguistic/code-switching issue, but regarding more 
general issues of word learning and comprehension, here's an analysis 
of related issues:

The literature on homonymy is most relevant. Check out the following 
(all w/ monolingual samples, I think). They do not paint a uniform 
(monochromatic?) picture but the general trend is that preschool 
children can interpret homonyms if reasonable contextual information 
supports the interpretation. However, even fairly old children can be 
"thrown" by low-frequency homonyms in sentential/story context (e.g., 
"The /har/ ran across the road" where"hair" is much more familiar & 
retrievable than "hare").
Here are some relevant references:
Children's resistance to homonymy: An experimental study of 
pseudohomonyms
  Journal of child language  [0305-0009]  Casenhiser, Devin  yr:2005  
vol:32  iss:2  pg:319  -343
Children's difficulty in learning homonyms
  Journal of child language  [0305-0009]  Doherty, Martin  yr:2004  
vol:31  iss:1  pg:203  -214
Possible explanations for children's literal interpretations of homonyms
  Journal of child language  [0305-0009]  Mazzocco, Michèle  yr:2003  
vol:30  iss:4  pg:879  -904
A comparison of homonym and novel word learning: The role of 
phonotactic probability and word frequency
  Journal of child language  [0305-0009]  Storkel, Holly  yr:2005  
vol:32  iss:4  pg:827  -853

The issue of so-called M.E. (implied by "reverse ME") is also relevant, 
and people often use "ME" in ways that aren't theoretically or 
empirically supported.  There is little evidence that children (of at 
least 2 years) generally assume that every class of referents has one 
label and each label refers to a unique (narrow) class of referents.
The evidence for this position is too extensive to summarize here, but 
for a slightly out-of-date review see:
	Deák, G. O. (2000). Chasing the fox of word learning: Why 
“constraints” fail to capture it. Developmental Review, 20, 29-80.
The references Roberta sent also are relevant; I won't rehash them.

Children readily show what Bill Merriman called the "disambiguation 
effect": when they hear a novel word in an ostensive context, all else 
being equal they tend to map it onto an unnamed referent rather than a 
familiar referent with a known (common) label. However, that tendency 
has a number of possible explanations, many not requiring an underlying 
assumption that things only have one legitimate label (or description). 
And, just to be completely clear, there is little or no support for the 
idea that children have a kind of representational inflexibility, such 
that they cannot represent an entity as belonging to only on category. 
To the contrary, for example, 2- to 4-year-olds readily and robustly 
produce (or accept) multiple appropriate labels for objects, people, or 
story characters:
	Clark, E. V. & Svaib, T. A. (1997). Speaker perspective and reference 
in young children. First Language.
	Deák, G. O. & Maratsos, M. (1998). On having complex representations 
of things: Preschoolers use multiple words for objects and people. 
Developmental Psychology, 34, 224-240.
	Deák, G. O., Yen, L., & Pettit, J. (2001). By any other name: When 
will preschoolers produce multiple labels for a referent? Journal of 
Child Language, 28, 787-804.

There are a few papers showing "true" ME effects (i.e., rejection of 
new word or replacement of old word (Merriman & Bowman, 1989; Deák et 
al 2001). Typically though the effects are small, but larger (it seems) 
when cognitive load is high (Liittschwager & Markman, 1994; Deák & 
Wagner, 2003). However, the Deak et al 2001 data show that 
rejection/replacement is an odd effect, and possibly an artifact: in 
short, right after learning a new word 3-/4-year-olds sometimes don't 
use or don't produce a familiar word for the same referent. However, 
there is nothing systematic about which word they don't produce--as if 
there is some maximum number of "legitimate" words they'll produce (in 
our studies, 2 to 3), and they choose from all the options...not at 
random, but not predictably either. The behavior looks like stochastic 
low-probability inhibition of lexical retrieval, possibly with specific 
items weighted by priming of that item. The implication is that this 
suppression doesn't say anything conceptually important about 
children's commitment to symbolic mappings (i.e., they're not committed 
to rejecting a word that they failed to produce on one occasion). I 
can't yet prove this account, but it's the best "fit" for the limited 
data available.
In a recent series of studies (under review; available on request) we 
showed that there is a weak, transitory ME-like effect for words, 
facts, and pictorial symbols, w/ no difference in magnitude across the 
three kinds of items. Notably, the effect disappears after 2 exposures 
to the novel information, and, again, it is obtained under a relatively 
high cognitive load (we haven't experimentally manipulated this yet). 
Best account: when learning new info. with many-to-many associative 
mappings, there is a slight, temporary inhibitory effect for 
overlapping mappings. In other words, Anderson's old fan effect is 
content-general, and accounts for "true" ME effects.

Two more brief notes (almost done!). First, there is converging data 
that children's pragmatic expectations influence the disambiguation 
effect: that is, children learn to expect grown-ups to use a restricted 
lexicon when talking to them (i.e., using shared words for a referent 
whenever possible (something like a Gricean principle of clarity). See, 
e.g., Diesendruck, Dev Psych 2005.

Second, Perner, Sprung, Doherty et al have proposed that young children 
have trouble generating alternate perspectives (i.e., dissimilar 
representations with positive truth-values wrt a referent), and this 
could account for ME effects. It is a clever account with some 
intriguing supporting evidence:
Perner, J., Strummer, S., Sprung, M., & Doherty, M. (2002). Theory of 
mind finds its Piagetian perspective: why alternative naming comes with 
understanding belief. Cognitive Development, 17, 1451–1472.
but is disconfirmed (as I read it) by other evidence, particularly:
Deák & Maratsos (1998)
Deák, Yen, & Pettit (2001)
Deák, G. O. & Enright, B. (in press). Choose and choose again: 
Appearance-reality errors and the logic of questioning. Developmental 
Science.

Best conclusions for now:
	Disambiguation is at least in large part due to pragmatic 
expectations. Exposure to a multilingual environment, and 
situation-specific information about an interlocutor's lexical 
knowledge, seems to "set" children's expectations about the lexicon 
that interlocutor will use. (We do not yet know how robust or precise 
this process is.)
	"True" rejection/replacement effects, including of homonyms, are 
generally small, transitory, content-general (i.e., not specific to 
novel words) and probably modulated by factors such as lack of 
contextual supporting information and cognitive load.
	Implications for the example: Given the apparent pragmatic basis of 
disambiguation effects, there is no reason why a child couldn't induce 
the expectation that speakers of L1 use different lexemes (i.e., 
phonological forms) than speakers of L2 for the conventional (i.e., 
common) forms in each language that refer to a single referent 
category. So hearing someone known to be an L2 speaker using the L1 
phonological form would violate this expectation. The child might 
respond in ways that reflect this expectancy violation. We know that 
children sometimes verbalize their doubt about a speaker's locution 
(whether or not they are committed to rejecting the locuation), and so 
a verbal statement of disbelief or rejection would not be unexpected.

Anyone who knows of other findings that contradict these conclusions, 
please tell me. Thanks!

On Mar 20, 2006, at 6:35 AM, Marnie Arkenberg wrote:

> Dear Info-CHILDES,
>    A colleague mentioned that her bilingual 3 year old refuses to 
> accept the notion that an object can have the same same in both 
> English and Hebrew. Can anyone guide me towards literature that 
> discusses this sort of thing?
> Cheers,
> Marnie Arkenberg
>
> Marnie E. Arkenberg, Ph.D.
> NIMH Postdoctoral Fellow
> 254F Baker Hall
> Department of Psychology
> Carnegie Mellon University
> Pittsburgh, PA 15213
> (412) 268-7986
>
>
>
Gedeon O. Deák, Ph.D.
Department of Cognitive Science
9500 Gilman Dr.								(858) 822-3352
University of California, San Diego				fax    (858) 534-1128
La Jolla, CA 92093-0515				   http://cogsci.ucsd.edu/~deak/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: text/enriched
Size: 9778 bytes
Desc: not available
URL: <http://listserv.linguistlist.org/pipermail/info-childes/attachments/20060322/7346f9e7/attachment.bin>