We
NS

RUTGERS

THE STATE UNIVERSITY
OF NEW JERSEY

DEPARTMENT OF COMPUTER SCIENCE «HILL CENTER FOR THE MATHEMATICAL SCIENCES
BUSCH CAMPUS + NEW BRUNSWICK «NEW JERSEY 08903 « (201) 932-3546/2001

November 7, 1977

Dr. Elliott Levinthal

Department of Genetics--IRL, S047
Stanford University Medical Center
Stanford, California 94305

Dear Elliott,

I received recently the Lesgold-Greeno proposal to SUMEX-AIM and a set of papers
that are related to the proposal. Here are my comments.

The main objective of this research is understanding of the reading comprehension
process and of relationships between certain cognitive parameters (e.g. short term
memory, word decoding processes) and certain reading inadequacies. The scientific
work, both empirical and theoretical seems to me well conceived and of high quality.
The use of ACT in the SUMEX-AIM framework is likely to facilitate the model building
aspects of this research. This is essentially work on information processing models
in cognitive psychology with some possible implications on applications to reading
instruction. The methodologies needed for theoretical development in this area strongly
overlap with work in AIM. The investigators face problems of interpretation and
diagnostic reasoning, and they wish to use AI techniques for building a theory which
accounts for the interaction between certain "surface processes" (i.e. word decoding)
and related "deep processes" which involve world knowledge and summary representations
of previous decodings. They may benefit in their work from the intellectual environ-
ment of AIM, and they may contribute to its enrichment. I see the main relevance to
AIM goals at the level of AI methodology and techniques, and concern with applications
of AI to scientific inquiry of psychological processes. Also, this is a good example
of dissemination of previous AIM work and increasing collaborations in the AIM com-
munity. I recommend approval of this proposal.

 

Singérely,

f

4

oe

 

Saul Amarel
Professor and
Chairman

SA/kb

omy