00100	Dear Lou:
00200	
00300	Would IBM be interested be interested in broadening its association
00400	with the outside speech recognition work through the mechanism of
00500	supporting some phase of out speech recognition work here at
00600	Stanford?
00700	
00800	I believe that this might be a good thing from IBM's point of view
00900	for the following reasons:
01000	
01100	1) It would bring a different perspective to bear on your problems
01200	with the enthusiasm and originality that fresh graduate students can
01300	offer.
01400	
01500	2) The Stanford work could be kept just enough separate from the rest
01600	of your speech work so that it could be discussed freely with others
01700	and act as the coin for gaining freer access into the ARPA-supported
01800	research community.
01900	
02000	3) Our point of view is much closer to yours than it is to the rest
02100	of ARPA group but at the same time the work that we are doing,
02200	particularly that on the pitch synchronous extraction of acoustic
02300	parameters, on the search for new acoustic correlates, the
02400	development of improved formant picking techniques that can be safely
02500	used in transition regions, and finally the development of adaptive
02600	learning techniques. all of this work would seem to nicely complement
02700	your in house work,
02800	
02900	4) My own long association if not active participation in speech
03000	work, first at the Bell Telephone Laboratories (1928-1946) where I
03100	was personally acquainted with all of the early workers in this
03200	field, and later at IBM (1949-1966) where I was more directly
03300	involved in the early faltering attempts to get started in this work,
03400	all of this background should be of value in preventing the needless
03500	repetition of old work.
03600	
03700	5) The time-sharing on-line graphics-oriented speech-research
03800	facilities here at Stanford are perhaps unexcelled, and enable
03900	certain things to be done that are otherwise quite difficult to do.
04000	
04100	Perhaps I should also add that I have been an IBM consultant off and
04200	on since my retirement in 1966 (but not in speech recognition
04300	research). I would be willing to consider a renewal of this
04400	relationship if this seemed desirable to provide better liaison
04500	between the two groups but this is not my major concern. My principal
04600	concern is to obtain some outside support for the graduate students
04700	in the Computer Science Department here at Stanford. The ARPA support
04800	that we have enjoyed has never been adequate and there is danger of
04900	its complete loss because of budget cuts and perhaps partly because of my
04950	 own outspoken lack of agreement with the ARPA point of view.
05100	
05200	If you want any additional information about me
05300	I suggest that you talk with some of the senior people in
05400	IBM research of if you want details of my current work specifically
05500	with Rex Dixon or Chuck Tappert.