Download as pdf or txt
Download as pdf or txt
You are on page 1of 16

History of Psychology Copyright 1998 by the Educational Publishing Foundation

1998, Vol. 1, No. 1,69-84 1093-4510/98/S3.00

THE LASHLEY-HULL DEBATE REVISITED


Darryl Bruce
Saint Mary's University and Emory University

N. Weidman (1994) claimed that "Karl Lashley and Clark Hull had a long and
unresolved controversy about the structure and function of the brain, its relationship
to the mind, and the use of machine metaphors to explain intelligence" (p. 162). The
record contained in published articles and unpublished correspondence indicates
otherwise. The clash was explicitly about continuity versus noncontinuity in
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

discrimination learning, stimulus generalization, and the development of quantitative


This document is copyrighted by the American Psychological Association or one of its allied publishers.

and mathematical psychological theory and its relation to neurophysiological data.


Weidman also contended that the subtext of the debate was whether heredity or
environment was more important in determining intelligence and behavior. This is
doubtful. It is more probable that the debate stemmed from Lashley's career-long
opposition to connectionism.

In a recent article, Weidman (1994) contrasted the views of Karl Spencer


Lashley (1890-1958) and Clark Leonard Hull (1884-1952), pictured in Figure 1.
With respect to the determinants of intelligence and behavior, she maintained that
Lashley gave more weight to heredity and Hull to the environment and that this
difference underlay their conflict.
On the surface, theirs was a debate about the structure and function of the brain, its
relationship to the mind, and the use of machine metaphors to explain intelligence.
But the relative importance of heredity or environment was the subtext of all their
exchanges. (Weidman, 1994, p. 162)
To reach these conclusions, Weidman described the differing ideas and biases of
the two men and considered their probable origins. For all practical purposes,
however, she ignored the debate itself, that is, as set forth in key publications of
Lashley and Hull (and also Kenneth Spence) as well as in relevant correspondence
by or about the two principals. In the present article I examine these sources. The

Darryl Bruce, Department of Psychology, Saint Mary's University, Halifax, Nova Scotia,
Canada, and Department of Psychology, Emory University.
Financial support has been provided by the Emory Cognition Project, Florida State University
Foundation, Marjorie Young Bell Fund of Mount Allison University, Saint Mary's University Senate
Research Committee, Social Sciences and Humanities Research Council of Canada, and Southern
Regional Education Board. Abe Amsel, Steen Larsen, Peter Milner, Jack Orbach, Mike Rashotte,
Roger Thomas, Gene Winograd, and Mike Zeiler criticized an earlier version of this article. I
appreciate their helpful recommendations even if I have not always heeded them. Letters and
unpublished documents cited in the text are from the following sources: Archives of the History of
American Psychology; Archives of George A. Smathers Libraries, University of Florida; McGill
University Archives; and Special Collections Department, Woodruff Library, Emory University. I am
indebted to the staffs of these institutions for their assistance in working with their collections.
This article was written while I was on sabbatical leave at Emory University. I thank the
members of the Emory Department of Psychology for their hospitality and for the use of the
department's resources.
Correspondence concerning this article should be addressed to Darryl Bruce, Department of
Psychology, Saint Mary's University, Halifax, Nova Scotia, Canada B3H 3C3. Electronic mail may
be sent to dbruce@huskyl.stmarys.ca.

69
70 BRUCE
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Karl Spencer Lashley (photo courtesy of Clark Leonard Hull (photo courtesy of
Robert A. Boakes and Frank A. Beach) Michael E. Rashotte and Ruth Hull Low)
Photo by Frank A. Beach

Figure 1. Karl Spencer Lashley and Clark Leonard Hull about the time their
disagreement began. Hull's picture was taken at the annual convention of the
American Psychological Association, 1936 (Amsel & Rashotte, 1984). Beach
recollected that he took the photo of Lashley at the same meeting (personal
communication to Robert A. Boakes, November 11, 1982). Photo of Lashley
copyright 1982 by Robert A. Boakes. Reproduced with permission. Photo of Hull
copyright 1977 by Ruth Hull Low. Reproduced with permission.

intent is twofold: to fill a gap in Weidman's treatment and to revisit the debate from
the perspective of some 50 years of hindsight.
To preview the present conclusions, we shall see that the written record of the
Lashley-Hull controversy furnishes quite a different picture from that set forth by
Weidman (1994). On the surface, their disagreement was not about the brain and
its relation to the mind, not about using machine metaphors to understand
intelligence. Instead, two scientific issues were at stake: continuity versus
noncontinuity in discrimination learning and stimulus generalization. The other
surface disagreement concerned theory. Hull emphasized formal mathematical
theorizing with little regard for the underlying neurophysiology of behavior.
Lashley's approach was just the opposite. As for the subtext, it was probably not
that Lashley and Hull differed over the extent to which heredity and environment
determined intelligence. Any subtext, such as it was, is more likely to have resided
in Lashley's long-standing and unrelenting opposition to connectionismhe
LASHLEY-HULL DEBATE REVISITED 71

instigated and fostered the debate because he perceived Hull and his associates as
connectionists.

Scientific Issues of the Debate


Continuity Versus Noncontinuity in Discrimination Learning
In acquiring the habit of consistently choosing one of two stimuli, does an
individual learn gradually to approach the positive stimulus and to avoid the
negative onethat is, does the learning occur in continuous incremental fashion?
Or is the discrimination acquired abruptly and completely in noncontinuous,
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

all-or-none fashion?
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Lashley had early on adopted a noncontinuity position based on his brain


ablation studies with rats. Here is how he put it in his 1929 monograph, Brain
Mechanisms and Intelligence:
There are many indications that... in the discrimination box, responses to position,
to alternation, or to cues from the experimenter's movements usually precede the
reactions to the light and represent attempted solutions which are within the rat's
customary range of activity....
The form of the learning curve . . . [in many cases] strongly suggests that the
actual association is formed very quickly and that both the practice preceding and
the errors following are irrelevant to the actual formation of the association. The
"all-or-nothing" action of the discrimination habit is very apparent in some studies
of the threshold of brightness discrimination which are now in progress.... [They
justify] interpreting the discrimination habit as a simple association acquired only
after a number of more familiar solutions have been tried unsuccessfully, (pp.
135-136)
Spence (1936, 1940) proposed an alternative to Lashley's views, a proposal to
which Hull (1943) later subscribed. It said in essence that all aspects of the positive
stimulus impinging on an animal's sensorium are gradually associated with the
response of choosing the stimulus; any association between all sensed aspects of
the negative stimulus and its choice is gradually weakened by consequent
nonreinforcement. Eventually, the difference between the two associative response
tendencies becomes sufficiently great to ensure that the positive stimulus will be
consistently selected. The key terms are all aspects and gradually. The learning
process is thus continuously cumulative. This constitutes the continuity or
Hull-Spence theory of discrimination learning.
Lashley's (1942a) reply was a more polished version of his inchoate ideas of
1929. An animal in a discrimination task, he asserted, responds to only a limited
subset of the elements that make up a stimulus; for example, size but not shape or
brightness. Which elements exactly is partly dependent on principles of attention
and partly on Gestalt principles of perceptual organization. A discriminative
response is associated only with those features of the stimulus to which the animal
selectively reacts and not with those it does not. In short, learning is not
continuously cumulative with respect to all aspects of a stimulus, hence the name
noncontinuity theory. Hypothesis-testing theory was another designation, the
implication being that the animal tries out various hypotheses in solving a
72 BRUCE

discrimination problem (e.g., Krechevsky, 1932). Thus Lashley's position had a


distinctively cognitive flavor.
The many subsequent experiments that attempted to decide between these two
theories need not concern us (for an especially clear review of the problem and
relevant literature, see Goldstein, Krantz, & Rains, 1965). Their upshot was that
they tended to favor continuity theory. On the other hand, Lashley's emphasis on
the process of attention was not misplaced, and the line of descent to blocking
experiments (e.g., Kamin, 1969) and the interpretation of conditioning in a
cognitive manner (e.g., Rescorla, 1988) may be drawn from Lashley.
Even Hull gave concession, if somewhat grudging, to the importance of
attention. In a letter (1942) to Spence he wrote
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

I expect with some confidence that sooner or later we must really do something
much beyond what we have yet done, in order to meet [Lashley's] objection based
on the subjective phenomenon of attention. I myself have no doubt whatever that
this is a genuine phenomenon in human adaptive behavior, and is not an illusion. I
have some tentative ideas as to the [stimulus-response] mechanism which may
mediate this.... I doubt some whether this mechanism is a very potent one, and
that it plays a very great role in ordinary animal learning.

Stimulus Generalization
Assume that an animal is conditioned to salivate in response to a light of a
particular brightness. The animal may then also salivate to a test light that is
brighter or dimmer than the conditioned stimulus. This is the phenomenon of
stimulus generalization. The animal may also display a stimulus generalization
gradient; that is, the amount of salivation may be directly proportional to how close
in brightness the test light is to the conditioned light. For Hull, stimulus
generalization was a first principle, a basic property of the nervous system. In fact,
he called it primary or physiological generalization (Hull, 1939).
Lashley's attack on the Hullian position appeared in an article entitled "The
Pavlovian Theory of Generalization" (Lashley & Wade, 1946).1 As Lashley saw it,
Hull's approach postulated two things: First, all elements of a stimulus are
gradually conditioned to a response (the continuity assumption); second, the
effects of training spread to other stimuli according to their similarity to the
training stimulus. Lashley's complaints were first, that learning is not continuous
(the same criticism he had leveled at Spence), and second, that the theory offered
no account of how generalization develops or what constitutes similarity among
stimuli.
From the position that learning is noncontinuous, Lashley claimed that much
stimulus generalization is generalization by default:
"Stimulus generalization" is generalization only in the sense of failure to note
distinguishing characteristics of the stimulus or to associate them with the
conditioned reaction. A definite attribute of the stimulus is "abstracted" and forms
the basis of reaction; other attributes are either not sensed at all or are disregarded.
So long as the effective attribute is present, the reaction is elicited as an all or none

'Marjorie Wade was a predoctoral fellow at Yerkes Laboratories of Primate Biology and is most
likely responsible for collecting some of the data reported in the article.
LASHLEY-HULL DEBATE REVISITED 73

function of that attribute. Other characteristics of the stimulus may be radically


changed without affecting the reaction. (Lashley & Wade, 1946, p. 81)
A gradient of generalization was another matter. That can arise, Lashley argued,
only if the individual has knowledge of the relevant stimulus dimension (size,
shape, color, etc.). Dimensions along which stimuli can vary may not exist in
advance for the subject in a generalization situation. If not, then a necessary initial
step is discrimination or differential training in which stimuli are varied along
whatever dimension of similarity the discrimination is to be based. "The
dimension arises as a result of attention to and comparison of two objects"
(Lashley & Wade, 1946, p. 78).
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Prior to its publication, Lashley sent Hull a copy of his critique and invited him
This document is copyrighted by the American Psychological Association or one of its allied publishers.

to respond (Lashley, 1945). Electing to do so, Hull (1947) patiently and


meticulously took account not only of Lashley and Wade's (1946) arguments but
also of comments that Lashley (1946b) had made in reaction to an advance copy of
the manuscript that Hull provided him. In his reply, Hull reviewed the extant data
on stimulus generalization and the conditions under which a generalization
gradient could be obtained and shaped. In particular, he pointed to evidence that
refuted Lashley's contention that a gradient can arise only after differential training
on two or more comparison stimuli. He even showed that Lashley and Wade's own
data were far from compelling on the point. In thus limiting his remarks to
Lashley's hypothesis that prior discrimination training is necessary to obtain a
generalization gradient, Hull appeared to have prevailed in the dispute. More
analytical research of the 1950s and 1960s, however, has borne out Lashley's
position (see Terrace, 1966, for a careful review of the issue). But absent such data,
Hull appeared to have had the edge in the controversy.
Hull also ignored Lashley's assertion that generalization is a failure of the
animal to associate a response with the distinguishing features of a stimulus, in
other words, that associative strength does not accrue to all features of a stimulus
but only to those that the animal pays attention tothe noncontinuity hypothesis.
Lashley lamented this in his letter to Hull of July 15,1946, in which he commented
on Hull's manuscript: "I had hoped that you would also discuss the continuity
theory." Then, describing his own noncontinuity position, he observed that "this
point of view lays great emphasis on attention. . . . We behaviorists threw out
attention in 1915 but it has come back for me as the central mystery of
psychology."
Hull would have none of the continuity-noncontinuity issue, however. Any of
a variety of reasons could have been persuasive: He may have thought that the
issue had already been decided in favor of continuity theory. We also know that he
did not consider attention a very significant problem. Then too, he may have
thought that Lashley, referring as he often did to Hull's ideas as conditioned reflex
theory, was still battling Pavlov. But whatever the reason, Hull steered clear of the
continuity question.
Having seen that the scientific content of the exchange between Lashley and
Hull centered around continuity in discrimination learning and stimulus generaliza-
tion, is there any room for Weidman's (1994) contention that "on the surface,
theirs was a debate about the structure and function of the brain, its relationship to
the mind" (p. 162)? In my judgment, not much. It is certainly true that Lashley was
antagonistic to the idea that learning consisted of point-to-point connections in the
74 BRUCE

cortex, especially between sensory inputs and motor outputs, and it also is true that
he concluded that the maze habit was not localized in the cortex. However, on
these issues Lashley's quarrel was with Pavlov, not Hull, for Hull held strongly to
no particular view of brain functioning. As Amsel and Rashotte (1984) have
pointed out:
Pavlov demonstrated how a program of behavioral research could provide the basis
for developing and testing hypotheses about speculative theoretical concepts. In
Pavlov's case, the hypotheses were about physiological terms and behavioral data
guided the development of a theory of cortical functioning. Theorists such as Hull
and Spence tested hypotheses about theoretical terms that had no necessary
physiological referents.2 (pp. 19-20)
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

So one should not expect to find anything about the brain and its functioning from
the hand of Hull in the documents pertaining to his disagreement with Lashley.
And indeed one does not. As we shall see, it was an omission that Lashley
continually harped on.
Debate About Psychological Theory
It is well known that mathematical and hypothetico-deductive theory reached
a peak in psychology with Hull (e.g., 1943). It is also well known that Lashley was
frequently critical of psychological theory. Can we say that Lashley was critical of
Hullian theory, which he was, simply because he was antagonistic to all theory?
Weidman (1994, pp. 163-164) strongly implied it. Hull himself probably thought
so. In a letter (1942) to Spence, he related that at a conference
I saw Lashley two or three times and told him that I would like to talk with him
about his long article on visual mechanism 3 .... At one of these brief meetings I
told him I thought I could now explain practically everything in that article. He
laughed in his characteristically hysterical manner and said that after I had done
that he would go into the laboratory and dig up a whole lot of new things which I
probably couldn't explain.... I am inclined to believe ... he has a thoroughgoing
distaste for any theory whatever, and would like very much to discredit every kind
of theory regardless.
Compelling as such stories may seem, the facts are that Lashley was far more
interested in making positive theoretical statements than is generally appreciated
(Bruce, 1994, 1996). Moreover, there were grounds for Lashley's criticism of
Hull's theorizing other than that he simply enjoyed being a critic.
What other grounds? One possibility is Weidman's (1994) claim that an
important surface characteristic of the debate concerned the use of machine
metaphors to interpret intelligence; Hull favored such metaphors and Lashley
opposed them. Examination of the published and unpublished records of the
controversy reveals that there was plenty of ink spilled in arguments about theory
but none about machine metaphors of intelligence.4

Initially, Hull and Spence differed in their concern for the underlying neurophysiology of
behavior; Hull offered physiological speculations and Spence avoided them. By late 1943, Spence
appeared to have persuaded Hull to adopt his more abstract approach. See Rashotte and Amsel (in
press) for a discussion of the subject.
3
The reference is probably to Lashley (1938b).
4
I do not deny that machine technology was an importance influence on Hull's thinking
LASHLEY-HULL DEBATE REVISITED 75

The actual disagreement between Lashley and Hull about psychological


theory is best described in their own words. First from Lashley's (1942a) rejoinder
to Spence's (1936) analysis of discrimination learning:
I am not unsympathetic with the type of quantitative theory which members of the
Pavlovian group have been trying to develop. Learning theory must ultimately be
cast in some such form, if it is to deal adequately with the established quantitative
relations between practice and improvement. But a survey of the phenomena of
learning suggests that other variables in addition to repetitive facilitation must be
taken into account as important factors in determining the rate of association in
discriminative learning. One of those variables is the ease with which figural
organization can be imposed upon physically independent items in the stimulus.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

For the present this variable cannot be quantified and, until it can be, quantitative
This document is copyrighted by the American Psychological Association or one of its allied publishers.

theories of learning are apt to prove abortive, (p. 264)


Here Lashley's complaint was mild: Mathematical theories of associative learning
were simply premature.
Four years later, however, Lashley's evaluation of the Hullian approach,
expressed in a footnote in Lashley and Wade's (1946) article, was startingly
caustic:
The spurious character of its quantitative and mathematical treatment of learning is
illustrated by the definition of its units of measurement, the hab and wat, in terms of
percentage of the practice required by a "standard" organism to reach the
physiological limit of learning. Such limits are completely devoid of meaning. The
physiological limit is in no case determinable and is a concept of questionable
value.... [One might] as well define the standard rat for discriminative learning by
length of tail as by performance in the maze. (p. 79)
Neither did he mince words about the gap between psychological theories of
stimulus generalization and neurophysiological knowledge:
No general laws descriptive of the processes by which a recognition of similarity is
reached have ever been formulated.... Both associationist and holistic systems
have fallen back upon the conception of a gradient of relations between perceptions
as the basis of similarity and generalization, but this gradient is in a purely
hypothetical medium having no substantial relation to the nervous system or to the
transmission of nervous excitation. It amounts to no more than a confession that the
basis of similarity and the mechanism of generalization are problems whose
solution depends upon the discovery of principles of nervous integration which are
as yet completely unknown, (p. 86)
Hull (1947) ignored the latter complaint in his published answer to Lashley's
attack. But the criticism of his quantitative mathematical treatment of learning was
another matter. On that count, the normally mild-mannered Hull came out swinging: "It
is difficult to believe that any behavior scientist would actively oppose the placing of the
behavior sciences on a secure quantitative, mathematical basis" (p. 132).
Lashley, seeing that statement in a copy of the manuscript of Hull's formal
reply that Hull had sent him for review, did not let it pass.
I certainly do not oppose the placing of behavior sciences on a secure quantitative
mathematical basis. But the making of consistent mathematical constructs and

(Rashotte & Amsel, in press) or that Lashley rejected machine analogies of brain functioning (Bruce,
1994). The objection is only that these different approaches, much less the use of machine metaphors
to understand intelligence (Weidman, 1994), were an explicit part of the debate.
76 BRUCE

determination of their fit to observable phenomena are different matters. The


implication of my (I fear) thoroughly nasty footnote is that the values or constants
necessary for your equations are indeterminable. (Lashley, 1946b)

Nettled even further by these remarks, Hull (1947) quoted them in a footnote
of his own in his published response (deleting only Lashley's admission that the
footnote may have been nasty). He italicized the last sentence for emphasis and
then stated that
we have in the italicized sentence quoted just above a confident and unequivocal,
though characteristically unsupported, statement that the scientific goal in question
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

is a human impossibility. Lashley thus deliberately puts himself on record. The next
This document is copyrighted by the American Psychological Association or one of its allied publishers.

few years should begin to show whether his intuition on this point is correct,
(p. 133)

Things were clearly getting out of hand, and both Lashley and Hull lowered
the heat in an exchange of letters in late 1949. First Lashley (1949) wrote to Hull,
commenting on a set of postulates distributed by Hull to interested parties on
October 4, 1949:
For many years I have been concerned with the problems you treat in your series of
postulates and have come out with a quite different set of concepts which seem
incompatible with your fundamental assumptions. I am not happy about this
situation but think I know how it has arisen. I am always trying to formulate
possible physiological mechanisms and to break down behavior into physiological
units. Consequently many phenomena take on a different relative importance and
definitions in molar terms fail to differentiate aspects of behavior which to me seem
essentially distinct. Some of the postulates also seem to imply improbable
physiological processes.

In a letter to Lashley, Hull (1949) brought the exchange to a close:


Upon further thought I am inclined to believe that there is a difference between our
approaches beyond that of the physiological versus molar. Perhaps it can be called,
"loaf versus half loaf." You must solve the problem of attention or else you will not
put forward your system. I, on the other hand, feeling much as you do in the matter
of attention, get along without it as well as possible until a solution is found,
meanwhile developing the rest of the system. I content myself, from this point of
view, with an incomplete loaf and you insist upon a whole loaf or nothing. Each
approach has its virtues.... Doubtless it is best for scientific progress for some to
choose one extreme, some to choose the other, and some to compromise in different
ways.

The clash between Lashley and Hull over psychological theory is at odds with
Weidman's (1994) perception that "on the surface, theirs was a debate about . . .
the use of machine metaphors to explain intelligence" (p. 162). A more accurate
depiction is that the explicit difference between the two protagonists was an
attraction to formal and quantitative mathematical theory with little concern for the
neurophysiology of behavior (Hull) and a search for the neurophysiological
mechanisms of behavior and the construction of neuropsychological theories
consistent with such mechanisms (Lashley).
LASHLEY-HULL DEBATE REVISITED 77

The Subtext of the Debate


Weidman's (1994) Hypothesis
What now about Weidman's (1994) statement that the subtext of the
Lashley-Hull controversy was the relative influence of heredity and environment
on intelligence and behavior? In support of her proposal she presented the case for
the hereditarianism of Lashley and the environmentalism of Hull and then
explained how such differences produced their divergent positions on brain
structure and function and the appeal of machine metaphors of intelligencewhat
she maintained were the surface features of the debate. But as we have seen, this is
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

not what the conflict was all about. Why then should we consider the relative
This document is copyrighted by the American Psychological Association or one of its allied publishers.

importance that Lashley and Hull attached to heredity and environment as the
subtext of their disagreement?
Perhaps, however, we can still draw causal links between Lashley's hereditary
leanings and Hull's supposed environmental emphasis5 and their respective stands
on the explicit issues of the debatecontinuity, stimulus generalization, and the
nature of psychological theory. In Hull's case, the linkage is nonexistent. Neither
his continuity position nor his proclivity to formal quantitative mathematical
theorizing appears to derive from his putative environmentalism. As for stimulus
generalization, Hull took that as a property of the nervous system, a nativist
assumption.
Concerning Lashley, the situation as a whole is equivocal. His noncontinuity
stance reflected an emphasis on inherited structure. Here are Lashley's (1949) own
words on the matter:

As a former geneticist, also, I am inclined to stress the importance of unlearned


organization. I believe that much of sensory and of motor patterning is a product of
the inherent structure of the nervous system, which results in a high degree of
selectivity with respect to what is learned. This . . . is the basis of my objection to
the continuity theory of learning.

On the other hand, there is nothing hereditarian about Lashley's other positions in
his debate with Hull. His insistence that psychological theories not be at odds with
neurophysiological facts has nothing to do with any hereditarian tendency. As for
stimulus generalization, Lashley maintained that the occurrence of a generaliza-
tion gradient depends on prior experience with the relevant stimulus dimension,
hardly a nativist position.
In sum, though Lashley and Hull may have differed in their assessment of the
relative influence of heredity and environment on intelligence and behavior, that
difference is not positively correlated with where they stood on all the issues that
they explicitly argued about. Hence the case for heredity versus environment as the
subtext of their exchange appears to have little foundation.

5
It may be noted that Hull did not exclude innate influences from his behavior theory. For
example, the third postulate of his system (Hull, 1943) states that unlearned stimulus-response
connections have the potential to evoke a hierarchy of responses. In general, Chapter 5 of Hull's
(1943) Principles of Behavior recognizes a role for innate tendencies in the behavior of organisms.
78 BRUCE

Another View
If there was a subtext, a more persuasive possibility is Lashley's long-standing
opposition to connectionism.6 It surfaced around the mid-1920s but became
prominent in 1929, first with the publication of his research monograph, Brain
Mechanisms and Intelligence (Lashley, 1929), and then with his address, "Basic
Neural Mechanisms in Behavior" (Lashley, 1930), delivered as President of the
American Psychological Association before the International Congress of Psychol-
ogy in New Haven, Connecticut, on September 4, 1929. The latter is especially
notable for its attack on Ivan Pavlov. Actually, Lashley never mentioned Pavlov,
only what he termed reflex theory. The essential fault of the theory as Lashley saw
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

it was that it postulated specific cortical connections between sensory receptors


This document is copyrighted by the American Psychological Association or one of its allied publishers.

and muscular effectors. As far as the maze habit was concerned, Lashley's brain
ablation studies with rats indicated no such specific connections or localization.
Pavlov, also an invited speaker at the Congress, heard Lashley's speech. To say
that it displeased him would be an understatement. His emotional reaction at the
meeting has been described by Windholz (1983), and his more considered
responsea 37-page article entitled "The Reply of a Physiologist to Psycholo-
gists" (Pavlov, 1932)has been thoroughly reviewed by Orbach (1982). The
particulars of these rejoinders need not detain us. Suffice it to say that Lashley was
unmoved by Pavlov's counterarguments.
Ever alert for signs of connectionism, Lashley was prone to attack wherever he
saw them. And see them he did in Hull-Spence theory. Spence's analysis of
discrimination learning? "Deduced from the fundamental assumptions of condi-
tioned-reflex theory" and an account of "discrimination learning in terms of the
familiar Pavlovian theory of differential conditioning," replied Lashley (1942a,
pp. 241-242). Hull's theory of stimulus generalization? Nothing more than "The
Pavlovian Theory of Generalization" (Lashley & Wade, 1946).
Yet the criticism began to take on a different cast. The focus shifted from
Pavlov to "the modern disciples of Pavlov"7 (Lashley & Wade, 1946, p. 72). In
particular, it was their abandonment of Pavlov's neurophysiology, indeed any
neurophysiology whatsoever, that rankled Lashley:
The neo-Pavlovian school has discarded the physiological assumptions of Pavlov's
theory but has retained a number of assumptions and interpretations of the
conditioned reflex which, although meaningful and reasonable in Pavlov's
formulation, become questionable and meaningless without his conceptions of
neurological activity. (Lashley & Wade, 1946, p. 72)
Even stronger evidence of Lashley's change of focus is his evaluation of
Pavlov's influence on psychology that he sent in a letter (1946a) to B. P. Babkin,
who at the time was writing a biography of Pavlov. Here is an excerpt from that
letter:
Pavlov believed, I am sure, that an understanding of behavior could be gained only
by study of the physiology of the brain, and that such studies should be as direct as

6
I use the term connectionism as Lashley would haveto refer to any theory that considers
learning to be mediated by connectionsneural, stimulus-response, or otherwiseand not, of
course, with reference to contemporary connectionist theory, notwithstanding that the latter stems
from the former.
'Lashley meant not only Hull and Spence but also B. F. Skinner.
LASHLEY-HULL DEBATE REVISITED 79

possible. He was primarily an experimenter and had little use for the construction of
conceptual systems, except as they could be supported step by step by direct
experimental evidence. But the history of the conditioned reflex in America has
been diametrically opposed to what was, I believe, his philosophy of science. He
pointed the way to fundamental investigation of the physiology of the brain. No one
in America has ever systematically followed that way or added significant factual
information to the fund that Pavlov provided....
As details of his work have become better known, more of his conceptions
have become embodied in speculative systems culminating in that of Hull. Such
systems from Watson to the present . . . have nothing in them of Pavlov's
experimental attitude.... My statement that the conditioned reflex was becoming
an obstacle to progress [Lashley, 1930] was based on the extent to which the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

concept was being used as a form of verbal magic to dismiss genuine problems
This document is copyrighted by the American Psychological Association or one of its allied publishers.

instead of submitting them to analytic investigation. It was a protest not against


Pavlov's point of view and experimental methods, but against the perversion of his
work from physiological fact to psychological mumbo-jumbo (e.g., Hull's irradia-
tion of conditioning along the physical dimensions of a stimulus series).
It should be appreciated that this assessment of Pavlov was written in the midst of
Lashley's dispute with Hull about stimulus generalization and quantitative
mathematical theorizing. So it can be no accident that Hullian theory was used on
two occasions to illustrate the level to which Pavlov's approach had degenerated,
at least in Lashley's eyes.
Lashley's opposition to connectionism and conditioned reflex theory certainly
did not end when the debate with Hull had concluded. As I have proposed
elsewhere (Bruce, 1996), it was probably at the root of the short shrift that Lashley
gave to Hebb's (1949) well-received neuropsychological theory. Hebb himself
noted this in a letter (1966) to Frank Beach:
[Lashley] didn't approve of the Organization of Behavior [italics added] at all when
I was writing it. (Perhaps I should say that when I first got interested in those ideas I
went to [Lashley] and suggested joint publication; he wasn't interested, and in fact
it took five months to get him to look over a draft outline I left with himbecause I
had told him it was a theory of connections, and that word was one he couldn't
stand.)
Indeed, even when Lashley put forth what appears to have been an essentially
connectionist theory of his own (Lashley, 1942b), he did not present it as such, nor
did he acknowledge any link between the theory he described in his final article
(Lashley, 1958) and Hebb's connectionist theory, despite their remarkable
similarity.
In short, Lashley was an uncompromising anticonnectionist throughout most
of his career. Moreover, it was a stance that led to many of his professional
disagreements, the one with Hull being no exception. But should it be regarded as
the subtext of their conflict? There are three caveats to consider, two in passing and
a third in detail.
First, because anticonnectionism was such an explicit part of Lashley's
theoretical makeup, it can be argued that it should not properly be regarded as the
latent subtext of his debate with Hull. I would not quarrel with that argument.
Second, Hull himself appears to have seen no hidden agenda in Lashley's
opposition to Hull-Spence theory. In a letter to Spence, Hull (1940) observed that
Lashley's
whole attitude seems to be one of serious conviction that our formulations have not
80 BRUCE

been adequate to deal with the rich complexity of the facts as he has learned them
from really many years of continuous investigation of this type of problem.
The third caveat derives from Weidman's (1994) position, and here we
encounter once more her conjecture that the subtext of the Lashley-Hull debate
was the relative importance that each attributed to heredity and environment as
influences on intelligence and behavior. Her claim, in essence, is that Lashley's
anticonnectionism actually arose out of his opposition to John B. Watson's
environmentalist!! and out of his conviction that intelligence is determined
primarily by heredity.
Weidman's (1994) reasoning requires close examination. Surveying the results
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

of Lashley's brain ablation experiments with rats during the 1920s, she observed
This document is copyrighted by the American Psychological Association or one of its allied publishers.

that even though the data themselves were much the same throughout that period,
Lashley moved in 1926 from a reflex theory interpretation of his findings to one
that argued against connectionism. Why? It is best to give Weidman's answer in
her own words:
This striking change in Lashley's science had its source in his decision to turn
against behaviorism. Lashley's rejection of the reflex theory coincided with
Watson's espousal [in 1924] of a radically environmentalist position in behavior-
ism, and I argue that the two events were linked. In seeking to distance himself
from what he considered Watson's extreme views about the power of the
environment to shape behavior, Lashley jettisoned the entire behaviorist system,
including, of course, the concept of the reflex character of all behavior. He also
removed all traces of environmentalism from his writings....
My argument is that his underlying allegiance was to a hereditarian belief in
the determination of intelligence, which was compatible with the behaviorism of
the early twenties, but not with behaviorism when it evolved into environmentalism
in the late twenties. Watson's "radical environmentalist" tract, which appeared in
1924, redefined behaviorism in the mid and late twenties, and lost followers like
Lashley, who believed it had ceased to be a useful theory, (p. 167)
Weidman's (1994) interpretation of Lashley's conversion from advocate to
opponent of reflex theory and connectionism is provocative and, though it appears
to square with some of the factsfor example, Lashley's well-known emphasis on
the importance of heredity in determining behavior and intelligencein too many
other places it takes on water.
The most troublesome matter perhaps is that Watson's turn to environmental-
ism occurred far earlier than Lashley's reinterpretation of his data and far earlier as
well than the radical environmentalism that Watson displayed in Behaviorism
(1924), the pivotal publication that Weidman believes caused Lashley's theoretical
change in 1926. Watson's move toward a more environmental account of behavior
was obvious in Psychology From the Standpoint of a Behaviorist, published in
1919. There one finds his views on the inherited basic emotionsfear, rage, and
love. Terming them reaction states, Watson described the stimuli and reactions
characteristic of these states in infants and declared that the "enlargement of the
range of stimuli capable of calling out emotional activity is responsible largely for
the complexity we see in the emotional life of the adult" (Watson, 1919, p. 211).
Then in 1920 came Watson and Rayner's article on the conditioning of emotional
reactions, which was partial payment on the environmental promissory note of
1919.
The other aspect of inherited behavior that concerned Watson in 1919 was
LASHLEY-HULL DEBATE REVISITED 81

instincts, and the chapter on that topic (Watson, 1919) also displays his
environmental tendencies. The extreme statement "Give me a dozen healthy
infants" (p. 82) of the 1924 book may not be there, but a host of other, more
considered indicators are: frequent cautions about reading too much instinct into
the behavior of adults simply because any instinctive reactions will have been
overlaid by habit; rejection of William James's position (that humans possess more
instincts than animals) on the grounds that humans have greater habit-forming
capacities than animals; discussion of a list of human instincts taken from
Thorndike and James with considerable dissent and questioning along the way as
to whether it is learning and not instinct that is involvedin brief, plenty of
evidence of Watson's call for a far more important role for experience and
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

environment in so-called human instincts. Was Lashley aware of this? Unquestion-


ably. In the acknowledgments section of the 1919 book, Watson thanked him for
having read the manuscript. So if Watson's environmentalism, which was
conspicuous in 1919, was instrumental in Lashley's taking an anticonnectionist
stand in 1926, why did Lashley delay so long in doing so?
Consider next Weidman's (1994) assertion that Lashley jettisoned the entire
behaviorist system. The problem with this claim is that Lashley never took it on
board in the first place (Bruce, 1986). In 1916, Watson and Lashley parted
company on their collaborative study of the conditioned reflex. Here is what
Lashley (1935) had to say about this in a letter to Ernest Hilgard:
Watson saw [the conditioned reflex] as a basis for a systematic psychology and was
not greatly concerned with the nature of the reaction itself. I got interested in the
physiology of the reaction and the attempt to trace conditioned reflex paths through
the nervous system started my program of cerebral work.
The point is that Lashley was a methodological behaviorist but never a subscriber
to behaviorism as a system. Indeed, he had a distaste for all systems of psychology
(see the earlier quotation from Lashley's letter to Babkin about Pavlov; Lashley,
1946a). So he was unlikely to have been perturbed by Watson's radical
environmentalism and even less likely to have changed his theoretical perception
of his findings because of it. In fact, I suspect that Lashley was actually
sympathetic to Watson's protest against the overappeal to instinct in explaining
complex human behavior.
Lashley did not eliminate all traces of environmentalism from his articles
because of Watson (Weidman, 1994). His empiricist account of generalization
gradients (Lashley & Wade, 1946) is but one example; another is his essay on
instinct (Lashley, 1938a) which, though it underscored the importance of innate
mechanisms of behavior, nevertheless reserved a place for learning and stressed
that environmental influences can change the character of instinctive responses
and their adequate stimuli. Lashley may have had an underlying allegiance to
heredity in interpreting intelligence but, like all well-trained geneticists, he was
well aware that genes and environment interact in the expression of behavior,
including intelligence.
To summarize, Lashley struck an anticonnectionist stance in the mid-1920s.
Why he did so then when he had hewed earlier to a different line even though there
may have been no fundamental change in his ablation research findings is a
significant question that awaits a convincing explanation. My own assessment, for
the moment at least, is that Lashley changed his mind for the same reason that
82 BRUCE

many scientists before and after him have done so: They simply come to perceive
their data differently. A classic example is Darwin's relinquishment of his belief in
the independent creation of species in favor of evolution. Whatever the reason, the
evidence indicates that Lashley's change to anticonnectionism had little if
anything to do with a desire to distance himself from Watson's radical behavior-
ism. Hence it may again be concluded that the subtext for the Lashley-Hull debate
was most likely not the different emphases that each may have given to heredity
and environment in accounting for intelligence and behavior.

Conclusions
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Weidman's (1994) assertion that on the surface the Lashley-Hull controversy


This document is copyrighted by the American Psychological Association or one of its allied publishers.

was "about the structure and function of the brain, its relationship to the mind, and
the use of machine metaphors to explain intelligence" (p. 162) is inaccurate. A
review of the published and unpublished records detailing the exchange shows that
it concerned the nature of discrimination learningcontinuous (Hull) or noncon-
tinuous (Lashley), stimulus generalizationan inherent property of the nervous
system (Hull) or dependent on past experience (Lashley), and a preference for
quantitative theory with no necessary neurophysiological referent (Hull) or for
psychological theory that squared with neurophysiological fact (Lashley).
Weidman's (1994) hypothesis that the subtext of the debate was the relative
importance of heredity (Lashley) and environment (Hull) as causal factors in
intelligence and behavior also is dubious. The actual surface features of the conflict
have no consistent correlation with any stress that the principals may have laid on
heredity and environment. Lashley was the primary instigator of the debate and the
general impetus for his attack was more likely his career-long aversion to
connectionism.
Weidman (1994) characterized the conflict as unresolved and intractable. That
assessment is correct. Still, it is revealing to look at the disagreement from the
personal perspectives of the two opponents. In a letter to Lashley, the genial Hull
(1946) showed that he regretted the exchange and why:
In conclusion let me say that I have a strong personal aversion to controversy and
polemics, and hitherto I have consistently refrained from becoming involved in any
even though repeated occasions have arisen. I am quite certain that I should not
have replied to your article had it come to my attention for the first time in
published form and with no special invitation to do so attached. It has involved a
considerable expenditure of energy, both on my part and on the part of my
laboratory staff, the dissipation of which could be ill afforded. However, there is no
use in weeping now over what is past.
For Lashley the controversy was business as usual, albeit business he had no
interest in continuing: "I do not want to engage in any more public controversy on
these matters, though our little exchange does seem to have had some value in
stimulating a number of experimental studies" (Lashley, 1949).
Lashley and Hull were at professional odds, but their personal relationship was
always friendly. It is notable, for example, that Hull was asked to contribute a
congratulatory letter to a book of such letters compiled in recognition of Lashley's
60th birthday:
May I take this occasion to recall some things from a number of years back. I
remember especially my contact with you during the summer of about 19278. . . .
LASHLEY-HULL DEBATE REVISITED 83

One evening we sat in your apartment and talked about psychological theory. Even
then we didn't agree very well, though I don't recall what we disagreed about.
(Hull, 1950)
Hull may have found the details irretrievable but, given that Lashley's opposition
to connectionism and conditioned reflex theory began to emerge in the 1920s, it is
not difficult to imagine what the general nature of their disagreement on that
summer evening would have been.

8
It was more likely 1928. That year Hull had an appointment as a visiting professor during the
summer session at the University of Chicago and Lashley was affiliated with the Behavior Research
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Fund of the Institute for Juvenile Research in Chicago.


This document is copyrighted by the American Psychological Association or one of its allied publishers.

References
Amsel, A., & Rashotte, M. E. (1984). Mechanisms of adaptive behavior: Clark L. Hull's
theoretical papers, with commentary. New York: Columbia University Press.
Bruce, D. (1986). Lashley's shift from bacteriology to neuropsychology, 1910-1917, and
the influence of Jennings, Watson, and Franz. Journal of the History of the Behavioral
Sciences, 22, 2744.
Bruce, D. (1994). Lashley and the problem of serial order. American Psychologist, 49,
93-103.
Bruce, D. (1996). Lashley, Hebb, connections, and criticisms. Canadian Psychology, 37,
129-136.
Goldstein, H., Krantz, D. L., & Rains, J. D. (Eds.). (1965). Controversial issues in
learning. New York: Appleton-Century-Crofts.
Hebb, D. O. (1949). The organization of behavior: A neuropsychological theory. New
York: Wiley.
Hebb, D. O. (1966, August 10). Letter to F. A. Beach. Hebb Papers, McGill University
Archives, Montreal, Quebec, Canada.
Hull, C. L. (1939). The problem of stimulus equivalence in behavior theory. Psychological
Review, 46, 9-30.
Hull, C. L. (1940, December 26). Letter to K. W. Spence. Spence Papers, Archives of the
History of American Psychology, Akron, OH.
Hull, C. L. (1942, July 2). Letter to K. W. Spence. Spence Papers, Archives of the History
of American Psychology, Akron, OH.
Hull, C. L. (1943). Principles of behavior: An introduction to behavior theory. New York:
Appleton-Century.
Hull, C. L. (1946, September 9). Letter to K. S. Lashley. Spence Papers, Archives of the
History of American Psychology, Akron, OH.
Hull, C. L. (1947). The problem of primary stimulus generalization. Psychological Review,
54, 120-134.
Hull, C. L. (1949, November 30). Letter to K. S. Lashley. Lashley Papers, Special
Collections Department, Woodruff Library, Emory University, Atlanta, GA.
Hull, C. L. (1950, April 27). Letter to K. S. Lashley. Lashley Papers, Archives of George A.
Smathers Libraries, University of Florida, Gainesville, FL.
Kamin, L. J. (1969). Predictability, surprise, attention and conditioning. In B. A. Campbell
& R. M. Church (Eds.), Punishment and aversive behavior (pp. 279-296). New York:
Appleton-Century-Crofts.
Krechevsky, I. (1932). "Hypotheses" versus "chance" in the pre-solution period in
sensory discrimination learning. University of California Publications in Psychology,
6, 27-44.
Lashley, K. S. (1929). Brain mechanisms and intelligence. Chicago: University of Chicago
Press.
84 BRUCE

Lashley, K. S. (1930). Basic neural mechanisms in behavior. Psychological Review, 37,


1-24.
Lashley, K. S. (1935, May 14). Letter to E. R. Hilgard. Hilgard Papers, Archives of the
History of American Psychology, Akron, OH.
Lashley, K. S. (1938a). Experimental analysis of instinctive behavior. Psychological
Review, 45, 445-471.
Lashley, K. S. (1938b). The mechanism of vision. XV. Preliminary studies of the rat's
capacity for detail vision. Journal of General Psychology, 18, 123-193.
Lashley, K. S. (1942a). An examination of the "continuity theory" as applied to
discriminative learning. Journal of General Psychology, 26, 241-265.
Lashley, K. S. (1942b). The problem of cerebral organization in vision. Biological
Symposia, 7, 301-322.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Lashley, K. S. (1945, August 14). Letter to C. L. Hull. Spence Papers, Archives of the
This document is copyrighted by the American Psychological Association or one of its allied publishers.

History of American Psychology, Akron, OH.


Lashley, K. S. (1946a, February 8). Letter to B. P. Babkin. Lashley Papers, Special
Collections Department, Woodruff Library, Emory University, Atlanta, GA.
Lashley, K. S. (1946b, July 15). Letter to C. L. Hull. Lashley Papers, Special Collections
Department, Woodruff Library, Emory University, Atlanta, GA.
Lashley, K. S. (1949, November 7). Letter to C. L. Hull. Lashley Papers, Special
Collections Department, Woodruff Library, Emory University, Atlanta, GA.
Lashley, K. S. (1958). Cerebral organization and behavior. In The brain and human
behavior: Proceedings of the Association for Research in Nervous and Mental Disease
(Vol. 36, pp. 1-18). Baltimore: Williams & Wilkins.
Lashley, K. S., & Wade, M. (1946). The Pavlovian theory of generalization. Psychological
Review, 53, 72-87.
Orbach, J. (Ed.). (1982). Neuropsychology after Lashley: Fifty years since the publication
of Brain Mechanisms and Intelligence. Hillsdale, NJ: Erlbaum.
Pavlov, I. (1932). The reply of a physiologist to psychologists. Psychological Review, 39,
91-127.
Rashotte, M. E., & Amsel, A. (in press). Clark L. Hull's behaviorism. In W. O'Donohue &
R. Kitchener (Eds.), Handbook of behaviorism. San Diego: Academic Press.
Rescorla, R. A. (1988). Pavlovian conditioning: It's not what you think it is. American
Psychologist, 43, 151-160.
Spence, K. W. (1936). The nature of discrimination learning in animals. Psychological
Review, 43, 427-449.
Spence, K. W. (1940). Continuous versus non-continuous interpretations of discrimination
learning. Psychological Review, 47, 271-288.
Terrace, H. S. (1966). Stimulus control. In W. K. Honig (Ed.), Operant behavior: Areas of
research and application (pp. 271-344). New York: Appleton-Century-Crofts.
Watson, J. B. (1919). Psychology from the standpoint of a behaviorist. Philadelphia:
Lippincott.
Watson, J. B. (1924). Behaviorism. New York: People's Institute.
Watson, J. B., & Rayner, R. (1920). Conditioned emotional reactions. Journal of
Experimental Psychology, 3, 1-14.
Weidman, N. (1994). Mental testing and machine intelligence: The Lashley-Hull debate.
Journal of the History of the Behavioral Sciences, 30, 162-180.
Windholz, G. (1983). Pavlov's position toward American behaviorism. Journal of the
History of the Behavioral Sciences, 19, 394-407.
Received December 16, 1996
Accepted February 14, 1997

You might also like