A preview of this full-text is provided by American Psychological Association.
Content available from Journal of Experimental Psychology: Animal Learning and Cognition
This content is subject to copyright. Terms and conditions apply.
INTRODUCTION
Developments in Associative Theory: A Tribute to the Contributions of
Robert A. Rescorla
Ruth M. Colwill
1
, Andrew R. Delamater
2
, and K. Matthew Lattal
3
1
Department of Cognitive, Linguistic, & Psychological Sciences, Brown University
2
Department of Psychology, Brooklyn College—City University of New York
3
Department of Behavioral Neuroscience, Oregon Health & Science University
The field of associative learning theory was forever changed by the contributions of Robert A. Rescorla.
He created an organizational structure that gave us a framework for thinking about the key questions
surrounding learning theory: what are the conditions that produce learning?, what is the content of that
learning?, and how is that learning expressed in performance? He gave us beautifully sophisticated ex-
perimental designs that tackled deep theoretical problems in experimentally clever and elegant ways.
And he left us with a collection of work that fundamentally altered the way we as a field think about ba-
sic learning processes. Few scientists have impacted their field in the way that Rescorla impacted animal
learning theory. In this paper, we introduce this special issue (Developments in Associative Theory: A
Tribute to Robert A. Rescorla) by considering some of the many ways in which Rescorla’s empirical
and theoretical contributions impacted learning theory over his almost 50-year career. We conclude by
identifying multiple fundamental issues we think he would have found especially fruitful to pursue as
we continue to move forward.
Keywords: Pavlovian conditioning, instrumental learning, Rescorla-Wagner model, extinction, learning-
performance problem
This special issue is devoted to the intellectual contributions of
Robert A. Rescorla (1940–2020) to the scientific study of basic
learning processes. Every so often in science a truly transformative
investigator comes along and directs the entire field with their
insights, methods, and discoveries. Along with Pavlov, Skinner,
Tolman, and Hull before him, Bob Rescorla was one of those inves-
tigators. Bob received his undergraduate training at Swarthmore
College where he was strongly influenced by the teachings of
Robert A. Rescorla (1940–2020), Christopher H. Browne
Distinguished Professor of Psychology - University of
Pennsylvania (Photograph courtesy of Shirley Steele)
Editor’s Note.This is an introduction to the special issue “Developments in
Associative Theory: A Tribute to Robert A. Rescorla.”Please see the
Table of Contents here: http://psycnet.apa.org/journals/xan/48/4/.—ARD
Andrew R. Delamater https://orcid.org/0000-0003-4638-7573
K. Matthew Lattal https://orcid.org/0000-0003-2885-449X
Each author contributed equally to this Introduction to the Special Issue,
and the order of authorship was determined alphabetically by last name.
The authors are especially indebted to Bob Rescorla for his guidance and
mentorship, his encouragement, and his friendship.
Correspondence concerning this article should be addressed to Ruth M.
Colwill, Department of Cognitive, Linguistic, & Psychological Sciences,
Brown University, 190 Thayer Street, Providence, RI 02912, United States, or
to Andrew R. Delamater, Department of Psychology, Brooklyn College—
City University of New York, 2900 Bedford Avenue, Brooklyn, NY 11210,
United States, or to K. Matthew Lattal, Department of Behavioral
Neuroscience, Oregon Health & Science University, 3181 Sam Jackson Road,
Portland, OR, 97239, United States. Email: ruth_colwill@brown.edu,or
andrewd@brooklyn.cuny.edu,orlattalm@ohsu.edu
245
Journal of Experimental Psychology:
Animal Learning and Cognition
©2022 American Psychological Association 2022, Vol. 48, No. 4, 245–264
ISSN: 2329-8456 https://doi.org/10.1037/xan0000344
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
Solomon Asch, Hans Wallach, and Henry Gleitman. Some of
Bob’s interests in Gestalt psychology concepts undoubtedly can be
traced to those early influences. Bob then entered graduate school
at the University of Pennsylvania to work with Dick Solomon.
Along with the other students in Solomon’s lab at that time, who
became extremely successful and influential in their own careers,
Bob especially stood out as unique. While a doctoral-level student,
Bob published two extremely influential articles in Psychological
Review that shaped how people study different aspects of Pavlovian
learning and its interactions with instrumental processes, and a 3rd
article in that same journal followed soon after he graduated from
Penn while he was a junior faculty member at Yale. These articles
addressed what constitutes an appropriate control procedure for the
study of Pavlovian learning and the importance of contingency,
over temporal contiguity, in learning (Rescorla, 1967b), how to
properly investigate Pavlovian conditioned inhibition (Rescorla,
1969b), and how to conceptualize the interactions between Pavlov-
ian and instrumental processes using Pavlovian-to-instrumental
transfer (PIT) procedures (Rescorla & Solomon, 1967). The signifi-
cance of these early articles is well known, but Bob was only just
beginning in his quest to fully investigate the nature of basic learn-
ing processes.
Bob is often regarded as the ultimate methodologist. His extremely
sharp mind was quick to identify the most essential features of a theo-
retical idea that would be worth testing. Most of the time, this meant
evaluating the most basic assumptions of a theoretical approach. His
contingency studies were a good example of this. At the time, temporal
contiguity was thought to be both necessary and sufficient for Pavlov-
ian learning to occur, but his work on contingency showed not only
that temporal contiguity was not sufficient for learning to emerge, but
that a common control procedure used at the time (the explicitly
unpaired procedure) itself generated convincing evidence for inhibitory
learning—a fact requiring investigators to completely rethink how
they go about assessing learning. Bob’s own approach was to think
about control procedures that were most appropriate for the theoretical
question of interest. So, for instance, if one wanted to discriminate
between excitatory and inhibitory learning processes, then a zero con-
tingency procedure would be most appropriate. On the other hand, if
one were most interested in determining whether CS-US pairings mat-
tered, then a paired versus unpaired comparison could be of value.
Ultimately, it was a theoretical question, and there was no single an-
swer as to what constituted the “best”control condition for all
situations.
Bob had an extraordinarily rigorous approach to conceptualiz-
ing other experimental control issues as well. He realized that a
wide variety of confounds could plague any experimenter’s task at
producing interpretable results, and this often led him to design
the most elegant experiments using within-subjects designs that
most investigators would agree would be best interpreted in one
way. A careful reading of virtually any of his nearly 200 empirical
articles reveals his clarity of thought and extreme care placed in
designing his experiments. This aspect of Bob’s approach set the
standard extremely high for the investigators’ability to assess im-
portant theoretical concepts controlling behavior. Most scientists
studying behavior today, at any level of analysis, would profit by
carefully studying his approach to control issues.
Although Bob is well-known for his methodological rigor, his
zeal for designing the “right”experiment was always in the service
of his interest in evaluating key theoretical positions. He, along
with Allan Wagner, of course, produced their famous Rescorla-
Wagner model (Rescorla & Wagner, 1972;Wagner & Rescorla,
1972) in the spirit of devising a common framework for under-
standing a wide variety of phenomena, some already discovered,
and others never previously evaluated or even imagined. The suc-
cess of that theoretical approach cannot be overstated, but it is fair
to say that Bob was never completely wedded to it. Ultimately, as
elaborated more below, he would be the first to recognize the mod-
el’s limitations, but he was interested in pushing its ideas as far as
he could with a willingness to adjust his thoughts on its most basic
assumptions based on what the data revealed. In this sense, he was
also the ultimate scholar who fully understood that science was a
process of discovery more than it was an opportunity to persuade
the scientific community toward his own way of thinking. He was
most unselfish in this way—the scientific study of basic learning
processes was not about him, it was about generating new knowl-
edge in search of the truth. But his own way of conceptualizing
the mechanisms of basic learning processes were often the most
influential, partly because of their elegance and parsimony.
One additional feature of Bob’s scientific genius was his ability
to “think outside the box.”Very often the scientific community
would focus its energy on understanding one, or another, aspect of
conditioning theory, but on several occasions throughout his career
Bob effectively redirected interest to problems that nobody had
previously considered. There are many good examples of this, and
some of these are described in more detail below (for example, his
focus on contiguity/contingency, within-compound conditioning
and Gestalt processes in compound conditioning, second-order
learning, binary/hierarchical associative structures in Pavlovian
and instrumental learning, the nature of the “prediction error”in
Pavlovian learning, etc.), but the fact that his focus on novel prob-
lems stimulated many other investigators to follow suit is a con-
crete measure of his widespread influence.
Bob’s specific strategy for approaching the problem of under-
standing basic learning processes was a three-pronged one
(Rescorla & Holland, 1976,1982). He thought it was essential to
address three key questions: (a) identifying the conditions critical
for learning to occur, (b) specifying the content, or associative
structures, of that learning, and (c) determining how such learning
ultimately was translated into observable performance. Although
Bob spent most of his research energy addressing the first two
questions, he also emphasized the importance of “the performance
problem,”and there are clear instances with respect to all three of
these questions where his work was ground-breaking. This organi-
zational structure has been adopted by former students and collab-
orators over the years and still provides a very clear framework for
understanding basic learning processes (e.g., Colwill, 2019;
Delamater & Lattal, 2014;Grau et al., 2022;Lattal, 2013). The
rest of this paper will explore in more detail some of Bob’s key con-
tributions within each of these, sometimes overlapping, domains.
What Are the Conditions Critical for Learning?
As noted above, one of the key problems in the study of associa-
tive learning has been uncovering the experimental conditions critical
for association formation. Bob Rescorla made significant advance-
ments on this problem that helped shape the research agenda for the
entire field and continues to inspire research to this day. As reviewed
in Nasser and Delamater (2016) many factors have been shown to
246 COLWILL, DELAMATER, AND LATTAL
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
have important influences on the development of excitatory associa-
tions between conditioned and unconditioned stimuli (CS, US,
respectively). Rescorla’s research, in one way or another, had a pro-
found impact on the study of most of these variables. Since the time
of Pavlov, CS-US temporal contiguity was recognized as important.
In addition, relative temporal contiguity, stimulus similarity, spatial
contiguity, CS intensity, US intensity, CS-US contingency, US sur-
prise, and CS-US “belongingness”have all been intensively studied.
Rescorla’s approach to many of these problems has always been
insightful and expansive. For instance, his early work expanded our
framework for thinking not only about the essential conditions for
the development of excitatory associations, but for inhibitory associa-
tions as well.
Conditioned Inhibition and Contingency Learning
Rescorla’s (1969b) important article on the study of Pavlovian
conditioned inhibition reviewed what was known at that time
regarding the conditions important for producing inhibitory condi-
tioning, and this article also established the methods by which in-
hibitory learning should be assessed (i.e., with retardation of
learning and summation test procedures). Moreover, the article
established a conceptual way of understanding how excitatory and
inhibitory learning processes might be related to one another, and,
within that context, provided a logical justification for assessing in-
hibitory learning with retardation and summation test procedures.
This work, as highly influential as it was, was conceptually con-
sistent with earlier important articles on the “truly random control”
procedure (Rescorla, 1967b,2000b) and the study of CS-US con-
tingency as a major determiner of learning (Rescorla, 1968,
1969a,1988c). His contingency studies and conceptual approach
to contingency learning helped launch the modern era of Pavlovian
conditioning research that emphasized information processing
ideas in the control of behavior. The beauty of Rescorla’s contin-
gency framework was that (a) it emphasized the important role of
“information”in the formation of associations, and (b) it provided
a unified framework for conceptualizing both excitatory and inhib-
itory learning processes. It is no coincidence that earlier frame-
works for conditioning emphasized excitatory learning and treated
conditioned inhibition as a separate process. That significantly
changed with Rescorla’s contingency work. Moreover, as noted
above, the framework logically pointed to the zero contingency
procedure as an appropriate control condition to assess learning,
but Rescorla recognized this issue more as a theoretical one
depending on what particular experimental question the investiga-
tor wished to address. Consequently, he was not wedded to this
control as appropriate for all circumstances.
Of course, Rescorla was quite aware of the problems with the
contingency framework. He never thought of it as a “theory”of
conditioning so much as an empirical description or “heuristic”of
the conditions critical for learning. Rescorla (1971) provided per-
haps the most compelling argument against the contingency
approach as a theory of learning. In that article, following Kamin
(1968), he used a blocking procedure to demonstrate that less learn-
ing would occur to a target CS conditioned in compound with
another CS that had been pretrained as a signal for the US. This
was compared with another group that, prior to compound condi-
tioning, had received the same number of CS and US presentations
in the first phase but with a zero contingency. Of additional interest,
though, was the observation that “superconditioning”was produced
during compound training to the target CS when it was conditioned
in compound along with a CS that had been trained as a conditioned
inhibitor in phase 1 (with a negative CS-US contingency). From the
point of view of the target CS (the added stimulus during compound
conditioning) the same CS-US contingency applied to all three
groups as the conditional probabilities of the USjCS and USjno CS
were equal in all three groups. In spite of this fact, drastically differ-
ent learning was obtained across them. This experiment, of course,
establishes that the “surprisingness”of the US was key.
The Rescorla-Wagner Model & Global Versus Local
Prediction Error Mechanisms
Results of the sort just described, together with those of Kamin
(1968) and Wagner (e.g., Wagner et al., 1968), led to the emer-
gence of the highly influential Rescorla-Wagner model (Rescorla
& Wagner, 1972;Wagner & Rescorla, 1972). It was through
Rescorla’s and Wagner’s ingenuity that they were able to explain
with a relatively straightforward, yet elegant, equation a wide vari-
ety of learning phenomena under the same theoretical umbrella
(Siegel & Allan, 1996). In short, the field now had a theory of con-
ditioning that could be used to make sense of many of the studies
investigating the critical conditions for learning to be established.
It made the simple statement that associative learning was a func-
tion of the surprisingness of the US, as well as the “salience”(e.g.,
as related to the intensity) of the CS and US. Both excitatory and
inhibitory learning could be understood, conceptually, within this
framework, and all of the “prediction error”driven learning phe-
nomena that ensued could also be readily understood. Of special
note were the observations that “protection from extinction”
(Rescorla, 2003b), “deepened extinction”(Rescorla, 2006a), and
“overexpectation”effects (Lattal & Nakajima, 1998;Rescorla,
2006b) could all be conceptualized from the same perspective. In
other words, the exciting possibility was that “extinction learning”
itself was not something that depended on the omission of reward
—rather, it depended on pairing the stimulus with a reduction in
the magnitude of an expected reward. Delamater and Westbrook
(2014) pointed out that a very understudied problem in the current
literature is understanding whether a common or several distinct
mechanisms might underlie different instances of inhibitory learn-
ing, all of which depend on negative prediction error processes.
The model also gave rise to a very explicit statement about the
conditions critical for generating inhibitory learning, as clearly
articulated by Mackintosh (1983), namely, that a stimulus accrues
negative associative strength when it is paired with a less than
expected US (be that no US or, merely, a reduced magnitude US).
To this day, the framework provided by the Rescorla-Wagner model
for understanding the conditions critical for learning to occur remains
extremely influential. Wasserman and Castro’s contribution to this spe-
cial issue further shows how this model relates to David Hume’s
theory of causation (Wasserman & Castro, 2022). The article by
Marchant and Chaigneau in this issue (Marchant & Chaigneau, 2022)
applies an adaptive filter model inspired by Rescorla-Wagner thinking
to problems of human categorization learning. While aspects of the
model are surely incomplete (e.g., Dickinson et al., 1976;Miller et al.,
1995;Pearce, 1994,2002), one reason why the model has enjoyed an
enduring success is that it receives support from underlying neurobio-
logical investigations of basic learning processes. In other words, the
A TRIBUTE TO ROBERT A. RESCORLA 247
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
brainseemstobestructuredinawaythatconformstosomeoftheba-
sic assumptions of the model (e.g., Iordanova et al., 2021;Waelti
et al., 2001). This point is particularly clear when looking at the neuro-
biological circuits responsible for Pavlovian conditioning. One key
assumption of the model is that fully anticipated USs lose their effec-
tiveness in supporting new learning. Kim et al. (1998) demonstrated
that fully anticipated USs are actually processed less effectively in the
cerebellar eyeblink conditioning circuit (i.e., they transmit the US sig-
nal very weakly to those brain regions where CS and US information
converge to support new learning). The reason for this reduced proc-
essing is also clear from the circuit. There is an inhibitory feedback
projection from the interpositus nucleus of the cerebellum to the infe-
rior olive and the CS more strongly engages this source of inhibition
as excitatory conditioning proceeds (Bengtsson et al., 2007;Bengtsson
et al., 2004;Hesslow & Ivarsson, 1996). Because the inferior olive
receives input from the US this source of inhibition ensures that the
US will lose its ability to be processed when expected. Kim et al.
(1998) demonstrated how this mechanism was causally responsible for
producing the blocking effect in this preparation, and Bengtsson et al.
(2007) demonstrated how this mechanism was responsible for extinc-
tion learning. These are just two examples, and Nasser and Delamater
(2016) review additional evidence from multiple learning paradigms
where negative feedback systems broadly play a key role in neural
plasticity (e.g., see Aggarwal et al., 2020;Fanselow, 1998;McNally
et al., 2011;Rasmussen et al., 2015;Steinberg et al., 2013). Certainly,
Schultz’s key observation that dopamine systems partly respond on the
basis of positive and negative prediction errors is entirely consistent
with the Rescorla-Wagner framework (e.g., Schultz, 1998;Tobler
et al., 2003).
Once again, though, Rescorla was his own worst critic. During
his later years, Bob returned to the question of what conditions
were critical for driving new associative learning. Whereas the
success of the Rescorla-Wagner model was widespread, other
learning models, notably Mackintosh (1975) and Pearce and Hall
(1980), offered different perspectives. One key difference was the
extent to which stimuli other than the target CS of interest contrib-
uted to the “prediction”that was part of the “prediction error”
computation. The Rescorla-Wagner model assumes that a global
error term is produced based on all stimuli present on the condi-
tioning trial, whereas earlier (e.g., Bush & Mosteller, 1951) and
contemporary (Mackintosh, 1975;Pearce & Hall, 1980) models
assumed that the prediction error computation is based only on pre-
dictions made by the individual target stimulus, in question. To
explain phenomena that clearly point to interactions among stimuli
conditioned in a compound, these other models had to resort to
changes in the associability of the stimulus rather than rely on
global prediction error mechanisms. With the key Rescorla-Wagner
model assumption that global prediction errors would be distributed
equally across all the stimuli present on a conditioning trial, Bob
evaluated that claim using what we take to be a most ingenious and
highly impactful experimental procedure he devised known as the
“compound test procedure”that clearly exemplifies his advanced
thought process. He realized that there were inherent problems with
attempting to measure the relative amounts of associative change to
stimuli that themselves differed in the level of responding they each
provoked. His compound test procedure, described in greater detail
below, was his solution to that problem. Interestingly, his work on
this particular problem revealed, contrary to Rescorla-Wagner
assumptions, that new learning was not equally distributed across
the elements of a compound conditioning trial. Instead, it appeared
as though stimuli whose individual predictions are most discrepant
with the actual trial outcome acquired more of the learning sup-
ported by that training trial. This finding points to the importance of
an “individual”prediction error term that might operate in concert
with a global error term (Rescorla, 2000a,2001c,2002a,2006c,
2008). Whereas Rescorla’s compound test procedure has generated
a wealth of new data relevant to the basic problem of understanding
what factors affect learning, his interpretation of the data continues
to enjoy lively debate in the literature (e.g., Chan et al., 2021;
Holmes et al., 2019;Honey et al., 2022;Jones et al., 2021;Spicer
et al., 2022;Uengoer et al., 2020).
Temporal Contiguity and Within-Compound
Conditioning Studies
Although the bulk of Bob’s work on the conditions critical for
learning to occur focused on questions centering around prediction
error concepts, he also performed important studies examining the
roles of other factors such as temporal contiguity, spatial contigu-
ity, and similarity. In the 1970s and 1980s Rescorla performed a
large number of studies exploring the role of associations between
the elements of a stimulus compound. The observation of so-called
“within-compound”conditioning (e.g., Rescorla, 1981a;Rescorla
& Cunningham, 1978b) pointed to the fact that associations can
form between CSs presented together, and that such associations
could complicate our analysis of learning that occurs between an
individual CS and the US. In one example of this, Rescorla and
Colwill (1983) noted that within-compound conditioning may
complicate interpretation of the “downshift unblocking”experi-
ment (e.g., Dickinson et al., 1976) that was taken as a challenge to
Rescorla-Wagner thinking. In that situation, unblocking was
obtained when a compound stimulus was paired with a single foot
shock US (AXþ) following a phase in which one of those stimuli
was pretrained as a signal for two foot shock US presentations
(Aþþ). Testing on stimulus X revealed that it had acquired signif-
icant conditioning (i.e., was “unblocked”). Under these conditions,
the Rescorla-Wagner model would anticipate that the X stimulus
should acquire inhibitory learning because it signals a downshift
in US magnitude. Rescorla and Colwill (1983) provided evidence
that responding to X in this procedure was partly determined by
the X-A within-compound association. That association would
counter the inhibitory learning to stimulus X predicted by the
Rescorla-Wagner model. Although such an account may not
explain the results of all of the downshift unblocking studies in the
literature that followed, the findings clearly point to the involve-
ment of learning processes that would not ordinarily have been
considered from more standard perspectives.
Another excellent example where such within-compound asso-
ciations could help explain “paradoxical”effects that challenged
prediction error concepts, was the “potentiation”experiment (e.g.,
Rusiniak et al., 1979). In this case, pairing a taste þodor stimulus
compound with illness resulted in greater odor aversion learning
compared with that produced by pairing an odor stimulus alone with
illness. Instead of observing overshadowing, the taste, apparently,
potentiated odor aversion learning. Durlach and Rescorla (1980),
however, provided evidence that odor-taste within-compound associ-
ations develop in this situation and could contribute to the manifest
odor aversion. Although we presently do not have all the answers
248 COLWILL, DELAMATER, AND LATTAL
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
about learning in these situations (e.g., LoLordo & Droungas, 1989),
the focus on within-compound associations illustrates another aspect
of Bob’s ability to think outside the box and direct the field to do the
same. Researchers had not really focused on what now seems like
the obvious possibility of two stimuli presented together associating
with one another or with the implications this might have for our
analysis of a wide variety of compound learning phenomena. But,
once again, Bob grasped its importance and saw it as a ripe avenue
for research.
In another especially clever study examining “within-compound”
associations, Bob asked the question of whether simultaneous or se-
quential presentations of two stimuli promoted stronger learning.
Temporal contiguity notions would support the view that optimal
learning should occur when two events are presented simultaneously,
as opposed to in sequence. However, years of Pavlovian conditioning
research indicated that simultaneous conditioning procedures tend
not to promote strong levels of conditioned responding (but see Bar-
netetal.,1991). Rescorla (1980b) suggested several factors that
could complicate analysis of the effectiveness of simultaneous pair-
ings in any conditioning procedure and devised an ingenious method
of distinguishing between the relative effectiveness of simultaneous
and sequential temporal pairings while holding constant any con-
founding factors arising from the fact that two stimuli are simultane-
ously presented during training but individually presented during a
test procedure. In particular, he gave the rats a compound taste solu-
tion to drink, AB, and immediately followed this with presentation of
a third taste solution, C. He then asked if the strength of the A-B
association was greater than the A-C association by devaluing the B
versus C solutions following AB-C training. Intake of taste A was
reduced more following devaluation of B relative to C, suggesting
that simultaneous AB pairing produced a stronger association than
had the sequential A-C pairing. This result is entirely consistent with
the notion that temporal contiguity plays a key role in association for-
mation. However, further investigations of such “within-compound”
associations point to the possibility that it represents a form of condi-
tioning fundamentally different from “ordinary”conditioning that
occurs when one event precedes another. For example, Higgins and
Rescorla (2004) showed that simultaneous taste-taste pairings gen-
erated learning that obeys principles quite distinct from those result-
ing from sequential taste-taste pairings. This theme is more fully
evaluated in the present special issue in the review article by Hall
that explores the role of extinction in flavor preference learning
(Hall, 2022). Bob’s own take on this was that within-compound
conditioning may entail the animal acquiring a kind of “configural”
representation of the AB compound and that this plays a strong role
in within-compound conditioning effects involving simultaneous
compound presentations. The theme of whether compound condi-
tioning involved “configural”or “elemental”processes is one that
interested Bob a great deal at various points throughout his career.
Although some of Rescorla’s empirical studies investigated tem-
poral contiguity issues, Bob was not especially focused on problems
involved in temporal control. However, Rescorla’s dissertation arti-
cle (Rescorla, 1967a) is probably one of the clearest examples of a
true “inhibition of delay”effect in Pavlovian fear conditioning in
that he documented strong evidence for inhibitory learning to the
early portions of a 30 s CS and excitatory learning confined to the
later portions of the same CS. We suspect he accepted Pavlov’s
general view that time within a stimulus could itself function as a
CS. Nonetheless, it is fair to say that Bob never fully developed his
approach to timing problems in conditioning. Still, he performed
experiments and entered into debates that were clearly relevant to
timing-based theories of Pavlovian conditioning.
One interesting situation where Bob contrasted differential predic-
tions between Rescorla-Wagner ideas and alternative timing-based
theories (e.g., Gibbon & Balsam, 1981) was the zero contingency
procedure. Bob’s approach has been criticized for being a “trials
based”approach to conditioning (e.g., Gallistel & Gibbon, 2000).
For instance, in the zero contingency procedure, Bob assumed that
presenting CSs and USs randomly with respect to one another over
the course of several conditioning sessions meant that the target CS
would occasionally be paired with the US in a background (the ex-
perimental context) that was also occasionally paired with the US on
its own. Following the Rescorla-Wagner model, zero contingency
training could be reduced to a blocking procedure in which the con-
text is thought to block conditioning to the target CS because it,
alone, predicts the US. One key question surrounding this view was
whether signaling those USs occurring in the intertrial interval (i.e.,
in the absence of the target CS) might impact learning to a target CS
that was, otherwise, randomly related to the US. Paula Durlach’sim-
portant work (while a doctoral-level student in Bob’slab)onthis
problem showed that, consistent with the Rescorla-Wagner theory,
signaling intertrial USs in a zero contingency procedure protected
the target CS from disruption and enabled it to enter into an associa-
tion with the US (Durlach, 1983). Grau and Rescorla (1984;also
Robbins & Rescorla, 1989) replicated this effect and went on to
demonstrate that the target CS and context mutually compete for
association with the US. The ITI signaling effect has also been
obtained in instrumental contingency experiments modeled after
Pavlovian studies (e.g., Dickinson & Charnock, 1985), and it has
posed a problem for alternative theories of conditioning that empha-
size a comparison process between the rates of reward both in the
presence and absence of the CS (e.g., Cooper et al., 1990;Gibbon &
Balsam, 1981).
Related to this issue is the question of whether learning develops at
all in normal zero contingency procedures and whether exposure to a
zero contingency procedure following positive contingency training
might eliminate such prior learning. Lindblom and Jenkins (1981)
reported results from a pigeon autoshaping task that often intrigued
Bob. In this task, pigeons were first trained on a positive keylight-
food contingency, and they developed conditioned keypeck responses
to the keylight CS that were eliminated by subsequent zero contin-
gency training. Of greater interest was that keypeck responding to the
keylight CS reemerged during an extinction test, suggesting that
although responding was reduced by the zero contingency procedure,
underlying learning was not erased (Lindblom & Jenkins, 1981). The
results, in principle, present problems for the Rescorla-Wagner theory
but could be explained from the perspective of a comparator theory if
one were to assume that context-based expectations of reward extin-
guish more rapidly than CS-based expectations during the nonrein-
forced test session. What is less well appreciated from this work,
however, is Lindblom and Jenkins (1981) additional finding, and later
replicated by Rescorla (1989), that such recovery fails to occur if the
CS had only been exposed to a zero contingency procedure prior to
the test phase (and had not been previously exposed to a positive con-
tingency). In other words, exposure to a zero contingency appears to
prevent initial learning from developing, but it may be sufficient to
maintain previous learning. Bob’s later work on the zero contingency
procedure revealed that subtle evidence for learning during the zero
A TRIBUTE TO ROBERT A. RESCORLA 249
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
contingency procedure can be obtained (Rescorla, 2000b), and it
reveals an enduring interest he had with the importance of the zero
contingency procedure along with its implications for a variety of
approaches to understanding conditioning. In the present issue, Austen
et al. (2022) explore the general issue of whether appetitive Pavlovian
learning in transgenic mice with a specific glutamate A1 receptor
knockout is best construed from the contrasting perspectives of trial-
based and reward rate based frameworks.
CS-US Similarity Studies
Other important studies of Bob’s related to the general question
of the critical conditions for learning involved his use of second-
order conditioning tasks to reveal the roles of spatial contiguity
and stimulus similarity (see Rescorla, 1980a). It is noteworthy that
from the time of Pavlov’s original studies, investigators generally
found second order conditioning to be a weak phenomenon. How-
ever, Rescorla and his students established robust second order
conditioning (e.g., Holland & Rescorla, 1975a;Rizley & Rescorla,
1972), and then the procedure was used extensively in the
Rescorla lab as a tool to address a wide variety of questions. For
instance, Rescorla and Furrow (1977),Rescorla and Gillan (1980),
and Rescorla and Cunningham (1979) all used second order condi-
tioning procedures in pigeon autoshaping tasks to investigate the
importance of similarity relations between first and second order
stimuli in promoting second order learning (see also, Grand et al.,
2007). In all cases, more substantial evidence for second order
learning was found when second order and first order CSs were
similar, either in terms of their physical features or their spatial
properties, compared with when they were dissimilar.
There are interesting implications of these findings. First, one
intensively studied topic since the 1970s has been the specializa-
tion of learning processes, as illustrated, for example, in the food
aversion paradigm. Garcia and Koelling’s (1966) famous finding
of “selective associations”became an intensely debated topic for
its implications for general process learning theory (e.g., Domjan,
2005,2015;Rozin & Kalat, 1971;Seligman, 1970). One type of
explanation offered is that spatiotemporal similarity between the
processing of foods and illness might play a role in association for-
mation (e.g., Krane & Wagner, 1975;Testa, 1975). Rescorla and
his students’studies on similarity clearly support such a view, at
least as a partial explanation of selective associations (LoLordo,
1979). A second important implication of these studies, though, is
how, in the first place, to conceptualize the importance of similar-
ity effects on association formation. One possibility is that a sepa-
rate process is sensitive to similarity relations between associated
events. However, Rescorla and Gillan (1980) and later Grand et al.
(2007) argued that similarity effects on association formation may
be reducible to other more primitive factors, such as stimulus in-
tensity and, what, ultimately, amounts to concurrent stimulus proc-
essing. This theme was more substantially developed in the
theorizing of Wagner (e.g., Brandon et al., 2000;Wagner, 1981;
Wagner, 2008;Wagner & Brandon, 1989,2001). As was true in
the study of stimulus selection (or cue competition effects, as they
are sometimes called) the effort was to reduce learning principles
to the concept of a prediction error mechanism. The Rescorla-
Wagner model gave us a common framework for implementing
such a mechanism. In the case of similarity effects, the question is
whether multiple mechanisms would be required or whether the
effects of interest could be reduced to a more primitive common
process (Rescorla & Gillan, 1980). Rescorla’s work on this prob-
lem revealed that some similarity effects could be reduced to
effects that can be described in terms of the concurrent processing
of events, an idea that is completely sympathetic with prediction
error models that emphasize how the processing of events is very
much dependent on experience. Additional research will be
required to fully address this question, but, as always, Rescorla’s
incisive research on these topics has surely led the way.
What Are the Contents of Learning?
Although Rescorla argued consistently for the importance of
asking three questions about learning, at times he may have con-
sidered the central problem for associative learning theory to be
defining what is associated with what (Rescorla, 1980a;Rescorla
& Holland, 1982). With regard to Pavlovian conditioning, his
research focused largely on identifying what aspects of the US are
associated with the CS where he and his students made enormous
strides using second-order conditioning; in instrumental learning,
his research focused on what associations develop between the dis-
criminative stimulus (S), the instrumental response (R), and the
instrumental outcome (O). Two of the most powerful techniques
that Rescorla applied to address these problems were to change the
value of the outcome after learning had occurred and to test for
outcome-specific transfer effects. By leveraging highly sophisti-
cated within-subjects designs, Rescorla and his colleagues used
evidence of selective outcome devaluation and outcome-specific
transfer to infer what the animal had or had not learned about the
US or O. An exquisite early example of Rescorla’s systematic and
elegant approach to the content of learning question is his 1980b
monograph Pavlovian Second-Order Conditioning. The impact of
the work that came out of his laboratory on within-compound
associations (Durlach & Rescorla, 1980;Rescorla, 1981a,1981b,
1982,1983;Rescorla & Colwill, 1983;Rescorla & Cunningham,
1978b;Rescorla & Freberg, 1978;Rescorla & Durlach, 1981;
Speers et al., 1980), sensory versus response encoding of the US
(Holland & Rescorla, 1975a,1975b;Nairne & Rescorla, 1981;
Rizley & Rescorla, 1972), binary versus hierarchical associations
in Pavlovian (Davidson et al., 1988;Davidson & Rescorla, 1986;
Rescorla, 1986a,1986b,1987a,1988b,1991a,1991b;Swartzen-
truber & Rescorla, 1994) and instrumental (Colwill & Rescorla,
1985b,1985c,1986,1988a,1988b,1990a,1990b;Rescorla,
1990a,1990b,1990c,1991a,1991b,1992a,1992b,1992c) condi-
tioning, and inhibitory S-R associations in extinction (Rescorla,
1993,1997a) has been broad and enduring.
Within-Event Learning
Rescorla advocated strongly for within-subjects designs and their
use in his laboratory has yielded some of the most elegant and com-
pelling demonstrations about the content of associative learning.
One of the most beautiful examples of this can be found in the
work by Rescorla and Cunningham (1978b) on within-compound
learning. In one experiment, rats were first exposed to two flavor
compounds, AB and CD. Following this, one element of one com-
pound, for example, B, was paired with a nausea-inducing agent
but another element of the other compound, for example, D, was
not. When the rats were given a choice between A and C, they
250 COLWILL, DELAMATER, AND LATTAL
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
preferentially consumed flavor C demonstrating that they had ini-
tially associated A with B and C with D. The internal controls for
any nonassociative effects were carefully enumerated. Also typical
of Rescorla’s experimental approach was the inclusion of at least
one replication of the basic effect within a paper. Here was no
exception. In a companion study, Rescorla and Cunningham
(1978b) replicated their within-compound learning effect using a
variation on the initial procedure. Another trademark feature of
work from the Rescorla lab was the systematic and exhaustive anal-
ysis of potential mechanisms and alternative interpretations. In this
case, they discussed two ways the AB within-compound association
might have operated to produce a change in the response to A when
B was devalued. First, during B’s devaluation, did B activate a rep-
resentation of A that was then paired directly with the aversive US?
Or second, during the choice test between A and C, did A activate a
representation of B which then affected its response? To date, there
seems to be no clear answer.
Second-Order Conditioning
For a number of years, Rescorla and his students focused their
efforts on describing the content of second-order conditioning
using pigeon autoshaping, rat appetitive and rat fear conditioning
procedures. Their approach was based on Rozeboom’s (1958)
logic that an association between two events, for example, S2 and
S1, could be revealed by changing the value of one event, S1, and
subsequently detecting a related change in the response to the other
event, S2. In one of the first series of experiments in the Rescorla
lab to ask the question does the animal learn about the sensory
properties of the outcome (S-S learning) or the response properties
of the outcome (S-R learning), Rizley and Rescorla (1972) used a
fear conditioning procedure in rats. They reported negligible
impact of extinction (simple nonreinforced presentations) of S1 on
the second-order conditioned response to S2. These results were
consistent with other rat fear and appetitive conditioning studies in
Rescorla’s lab (Holland & Rescorla, 1975a;Rescorla, 1973b,
1977) and suggested that a representation of the outcome (S1)
does not participate in S2-S1 learning, consistent with S-R learn-
ing. Bob was fond of using these findings in his learning courses
to advocate against some psychoanalytic approaches that focused
on searching for the origin of a fear or anxiety in their clients. In a
later series of experiments, Nairne and Rescorla (1981) revisited
this issue using extinction to change the value of S1 after S2-S1
second-order pairings in a pigeon autoshaping procedure. They too
found that the second order stimulus, a visual S2, was impervious
to a change in the value of S1 when S1 was an auditory stimulus.
However, they discovered that when S1 was also visual, the visual
S2 was sensitive to a change in the value of its S1. This result indi-
cates that when S2 and S1 are in the same modality, encoding of
the sensory features of S1 is encouraged. Such thinking was con-
sistent with Rescorla (1980a) where a common modality between
S2 and S1 in a fear conditioning procedure with rats resulted in
responding to S2 tracking changes in the response to S1. Such S-S
learning was also encouraged if S2 and S1 were presented simulta-
neously while S-R learning was encouraged when the two were
presented sequentially (Rescorla, 1980a). A few years later, Col-
will and Rescorla (1985a) using second-order conditioning in
pigeon autoshaping reported that S-S learning was not affected by
the amount of S2-S1 training, partial reinforcement of S2, or the
length of the S2-S1 interval, all variables that had been suggested
in other research to encourage S-R learning (e.g., Adams & Dick-
inson, 1981;Chen & Amsel, 1980;Haas et al., 1970;Morgan,
1974,1979;Rescorla, 1977). Studies of second-order conditioning
in Rescorla’s lab indicate that the answer to the question Is
learning about the outcome S-S or S-R? is It depends and reveal
some of the parameters that affect the extent of learning about
the stimulus properties of the outcome. What is associated with
what in second-order conditioning continues to be an area of
active investigation with interesting questions emerging in neu-
robiological research (e.g., Fam et al., 2022;Gostolupce et al.,
2022;Holmes et al., 2014).
Modulatory Processes
Rescorla was not only a clever architect of experimental design
but also a deep thinker who was willing to challenge the orthodox
associative framework if that’s where the data led. Along with but
independently of Herb Jenkins and Peter Holland in the 1980s, he
helped shape thinking about a far-reaching limitation of the stand-
ard associative learning framework, namely, that binary associa-
tions of various kinds alone described learning. In his studies of
modulatory processes, he reconceptualized conditioned inhibition
as the opposite of facilitation, not of conditioned excitation as had
been the case for the Rescorla-Wagner model, and this pointed to
a new type of function mediating learning. For example, he dem-
onstrated that whereas simple nonreinforcement extinguished an
excitor, it had no detrimental impact on a facilitator or an inhibitor,
and might even improve their ability to transfer to new targets
(e.g., Zimmer-Hart & Rescorla, 1974). Further, he showed that
extinction of a facilitator occurred when it was trained as a con-
ditioned inhibitor. Rescorla hypothesized a new associative func-
tion that went beyond what can be described by simple binary
associations alone, namely, that facilitators and inhibitors modu-
lated the threshold for activation of the US and thus would trans-
fer, a view and prediction that conflicted with Holland’s account
that modulators operated on the association between the trained
CS and the US and thus should not transfer. Two articles from
this special issue address these issues. A key finding from Stein-
feld and Bouton (2022) is that instrumental inhibitors transferred
across responses trained with different outcomes. Lovibond et al.
(2022) show that inhibition is not reversed by simple extinction,
but it is reversed by explicitly removing the inhibitor’s ability to
modulate in human contingency learning. Bob believed quite
firmly that alternative theories or models were essential to the
health of a discipline and even if he thought they were wrong, he
leveraged them to probe for the truth.
Another example of Bob’s quest for the truth can be found in his
articles on the nature of the representation of a compound stimulus.
In one series, he developed the unique cue hypothesis to explain
how the animals solved complex Pavlovian associations such as
negative patterning and conditional discriminations while preserv-
ing an elemental model approach. In the case of negative patterning
or the XOR problem, he argued that the lack of responding to the
nonreinforced compound AB when the elements A and B were sep-
arately reinforced could be explained by an inhibitory unique cue
that was present on the AB trials. Rescorla et al. (1985) examined
the unique cue in two configural discriminations cleverly leveraging
second-order conditioning. For example, pigeons were trained on
A TRIBUTE TO ROBERT A. RESCORLA 251
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
an Aþ,Bþ,ABdiscrimination. Then, X was paired with A and
Y with B. When XY was tested in compound, the pattern of
responding was similar to AB. They interpreted their results as
showing that “any unique cue controlling responding to their com-
pounds does not depend on the physical presence of the component
stimuli. Instead, the unique stimulus appears to arise from the joint
activation of memory representations.”Bob was also interested in
distinguishing configural representations from hierarchical associa-
tions not only in Pavlovian discrimination learning but also in
instrumental learning. Two articles in the present special issue take
up these issues of configural learning (Gupta et al., 2022;O’Donog-
hue et al., 2022).
Associations in Instrumental Learning
Instrumental learning has proven to be one of the most fertile
areas for creative ways of thinking about what is learned, and Bob
indisputably left an indelible mark on this area. While still a gradu-
ate student at UPenn, Bob coauthored an article with his advisor
(Rescorla & Solomon, 1967) that curated evidence for the supple-
mentary role of a Pavlovian S-O association to motivate perform-
ance. Their two process theory (Pavlovian S-O and instrumental
S-R) has had considerable influence in our understanding of the
mediation of instrumental responses by Pavlovian cues. But his
later work with Colwill contributed to a seismic shift in the analy-
sis of instrumental learning. Building on work by Adams and
Dickinson (1981), Colwill and Rescorla (Colwill & Rescorla,
1985b,1985c,1986,1988a,1988b,1990a;Rescorla & Colwill,
1989) used elegant within-subjects designs coupled with two
powerful behavioral techniques (outcome devaluation and out-
come-specific transfer) to compile compelling evidence regarding
the insufficiencies of classical S-R theory (e.g., Hull, 1943) and its
two process cousins (Rescorla & Solomon, 1967;Trapold &
Overmier, 1972) that had dominated explanations of instrumental
learning. In their initial studies using two responses and two out-
comes, they proposed that the outcome was associated with its
response (an R-O association), a view in line with that recently
promoted by Dickinson and Mackintosh (Dickinson, 1980;Mack-
intosh & Dickinson, 1979) and advocated by others (Bolles, 1972;
Tolman, 1932). Furthermore, they found negligible evidence to
support the orthodox view of the outcome as a catalyst for the
instrumental S-R association. Indeed, in their hands, extended
training simply strengthened the R-O association (see also Colwill
& Triola, 2002;Garr et al., 2020;Garr et al., 2021) and residual
responding that was sometimes observed following outcome
devaluation, although consistent with S-R learning, could be
attributed to other sources such as incomplete outcome devalua-
tion (Colwill & Rescorla, 1990a). Subsequent work using the
transfer test in which responses trained with different outcomes
(e.g., R1-O1 and R2-O2) were tested in the presence of discrimi-
native stimuli trained with another response and different out-
comes (e.g., S1: R3-O1 and S2: R3-O2) revealed evidence for S-O
learning. Here, in a choice extinction test, S1 selectively promoted
R1 and S2 selectively promoted R2 confirming R-O associations
and S-O associations. Note that in this and many other cases, trans-
fer is made on the basis of the S-O association thought to be
acquired by the instrumental discriminative stimulus. Whether or
not that S-O association is equivalent to the S-O association learned
in Pavlovian conditioning and revealed in outcome-specific PIT
was an issue that Bob thought required additional experimental
work.
Soon after their initial work identifying binary S-O and R-O
associations in instrumental conditioning, Colwill and Rescorla
(1990b) published three converging lines of evidence to suggest
that instrumental learning might have a hierarchical architecture,
an idea that was first alluded to by Skinner (1938). In one study,
conceptually related to Asratyan’s (1965) switching design, Col-
will and Rescorla (1990b) used the outcome devaluation procedure
to interrogate what rats had learned in an instrumental switching
task. The design involved two stimuli, two responses, and two out-
comes. Rats were trained so that R1 and R2 earned O1 and O2,
respectively, during S1 and O2 and O1, respectively, during S2.
Then, one outcome (e.g., O1) was devalued but the other (e.g.,
O2) was not. During a choice extinction test, R1 was depressed
relative to R2 during S1 but R2 was depressed relative to R1 dur-
ing S2. This pattern of a reversal in response depression between
S1 and S2 was not anticipated by any binary model or by any com-
bination of binary associations. Instead, it lent support to the idea
that the rats had encoded the three-term contingency. However,
rather than adopting a modulatory (threshold shifting or gating)
account analogous to that used to explain facilitation and inhibi-
tion, they proposed an associative model in which the R-O relation
was treated as a unit to be associated with its S (see also Rescorla,
1990a,1991b).
The groundwork laid by these studies shaped the questions that
were to follow. What is the nature of the S-O association? How
does training with multiple outcomes affect a target R-O or S-O
association? What are the conditions for extinguishing R-O and S-
O associations? What role might an O-R association play?
Rescorla addressed these questions, as always, in clever, well-
designed experiments. He was convinced by the data that a dis-
criminative stimulus was not equivalent to a Pavlovian CS, that S-
O and R-O associations persisted through a variety of decremental
operations, and that the S did not influence the ease with which
the outcome representation could be accessed. For example,
regarding the latter point, he amplified the empirical basis for an
earlier claim (Colwill & Rescorla, 1990a) that transfer does not
depend on the value of the mediating outcome (Rescorla, 1990b,
1994). This led to the conclusion that S activates a representation
of O which in turn retrieves a representation of R through the use
of the R-O association in a backward direction. As a result, S
would promote that R to the extent that R had a valued outcome
but that valued outcome need not be the same as the mediating
outcome. Rescorla (1992c) went to some lengths designing more
brilliant experiments that could discriminate between instrumental
learning of R-O versus O-R associations (Trapold & Overmier,
1972). His results showed that when rats lever pressed for O1 in
the presence of a stimulus that otherwise signaled O2, the rats’
behavior, ultimately, was controlled by the R-O1 and not the O2-
R association.
The persistence of S-O and R-O associations through extinction
and other decremental operations was robustly observed in instru-
mental learning studies (Colwill, 1994;Rescorla, 1991a,1992a,
1995). Analogous persistence of CS-US associations was also a
routine finding in Rescorla’s (1996a) studies and those of Delam-
ater (1996; though see also Alarcón & Delamater, 2019;Delamater
et al., 2017). Many of the experiments used outcome-mediated
transfer to detect the presence of the original learning. In this issue,
252 COLWILL, DELAMATER, AND LATTAL
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
Crimmins et al. (2022) revisit the question of whether the R-O asso-
ciation is indeed impervious to change when a random or zero con-
tingency is used as the decremental operation. Their studies use
spatially separated levers or a bidirectional lever in combination
with outcome devaluation through selective satiation to detect the
status of the decremented and nondecremented R-O associations.
Outcome devaluation and transfer mediated by a shared outcome
are two powerful techniques for probing the structure of instrumen-
tal learning and Rescorla leveraged them to the full. They are also
routinely incorporated into research outside of the Rescorla lab
(e.g., Alarcón & Delamater, 2019;Cartoni et al., 2016;Colwill,
1994,2007;Colwill & Delamater, 1995;Delamater et al., 2017;
Hogarth et al., 2007;Holmes et al., 2010;Laurent & Balleine,
2015;Kosaki & Dickinson, 2010) and have been used to advance
our understanding of the neural basis of instrumental learning (e.g.,
Balleine, 2019;Corbit & Balleine, 2005,2011;Coutureau & Kill-
cross, 2003;Killcross & Coutureau, 2003). Moreover, some studies
have attempted to understand the neural basis of the techniques
themselves (e.g., Laurent et al., 2012;Laurent et al., 2021;Lichten-
berg et al., 2017;). There is no doubt that multiple types of associa-
tions, S-O, R-O, O-R, S-R, and S-(R-O), describe what is learned
and contribute to instrumental performance. However, it was not
until Bob started his studies of extinction that he compiled evidence
for learning involving an S-R association in instrumental condition-
ing. Assessing the status of S-R associations with more direct tests,
as opposed to inferring their contributions by default, remains an
issue in need of further experimental development.
What Is Learned in Extinction?
When asked for advice, Bob was known for encouraging future
generations of researchers to find problems that others were not
working on. He himself would reach a point where, after having
seemingly exhausted possibilities for advancement in an area, he
would move on to a new challenge or perhaps briefly revisit an
old one. One of the last questions he addressed about the content
of learning is what is learned in extinction. Once again, his work
was characterized by incisive thinking and elegant design. While
he recognized the complexity of the problem and the improbable
likelihood of a single mechanism underlying extinction, he com-
piled persuasive evidence for an outcome-independent decremen-
tal process. He saw this as an inevitable solution to explain the
decrease in both Pavlovian conditioned responses and instrumental
responses during extinction, given the body of evidence that out-
come associations (S-O and R-O) persisted through a variety of
extinction procedures (Colwill, 1994;Delamater, 1996;Rescorla,
1991a,1992a,1995,1996a). In a review of extinction, Rescorla
(2001b) draws our attention to what the animal’s experience is in
transitioning to a simple extinction procedure. For example, in
Pavlovian conditioning, the previously reinforced stimulus is pre-
sented and no US follows but the animal makes a response. That
response will include an overt component, the CR that has been
measured, and an emotional component, frustration, at the omis-
sion of an expected US. Similar consequences occur in instrumen-
tal conditioning when the previously rewarded response is made
and the expected outcome is withheld.
Rescorla (2001b) proposed that the negative emotional conse-
quences elicited by the omission of the expected outcome become
associated with the Pavlovian and instrumental stimuli and
responses, somewhat analogous to the S-R association that devel-
ops in some second-order conditioning procedures. Concurrently,
he argued that the effectiveness of extinction was positively corre-
lated with the number of overt responses made which strongly
suggests some learning involving the response itself. Bob fully
appreciated the difficulty in providing direct evidence for the pres-
ence of S-R associations, but using his signature within-subjects
design and the outcome-mediated transfer technique, Rescorla
(1993,1997a) obtained converging evidence to support the contri-
bution of an inhibitory S-R association proposed by Colwill
(1991) to the extinction of both Pavlovian and instrumental
responses. Incidentally, a comparison of his two articles on this
topic provides a nice example of how he would replicate and
strengthen his original results. In one of the Rescorla (1993)
experiments, he trained two discriminative stimuli with a common
response that earned pellets. This established S1-O1 and S2-O1
associations. Then, he extinguished R1 in S1 and R2 in S2. Both
R1 and R2 had been trained with O1. In the choice test with both
R1 and R2 available, presentations of S1 preferentially augmented
R2 and presentations of S2 preferentially augmented R1. This dif-
ference is consistent with the development of an inhibitory S-R
association between S1 and R1 and S2 and R2 during their extinc-
tion (Rescorla, 1993). In the follow up article, Rescorla (1997a)
augmented that simple design to exclude the possibility that the
earlier result was a renewal effect brought about by a change in
context (e.g., Bouton & Bolles, 1979) and to make the point that
the strength of the inhibitory S-R association depended on the
number of times the response was made in extinction. The design
was constructed so that the conditions of extinction and testing
were the same for both responses but the relative rates of the two
responses in the stimuli were manipulated by exploiting the degree
of outcome-mediated transfer thus allowing for variations in the
strength of potential inhibitory S-R associations. To accomplish
this, he established two different instrumental S-O associations
using a common response, S1-O1 and S2-O2. Then, he trained
two new responses, R1 and R2, with different outcomes to estab-
lish R1-O1 and R2-O2 associations. In the extinction phase, R1
and R2 were both available and S1 and S2 were occasionally pre-
sented. Outcome-specific transfer would ensure that R1 occurred
more often than R2 during S1 but that the reverse would happen in
S2. R1 and R2 were then retrained in the absence of S1 and S2 but
with the outcomes reversed (i.e., R1-O2 and R2-O1). Finally, R1
and R2 were tested with S1 and S2. During this test, the more a
stimulus had evoked its response during extinction, the less it now
promoted that response (e.g., R1 would occur more frequently in
S2 than in S1). This study nicely illustrates his ingenuity in not
only conceptualizing a problem, but also his expertise in devising
multiple ways of testing the basic ideas leading to a highly com-
pelling conclusion.
How Is Learning Translated Into Performance?
The problem of performance—how is learning translated into
behavioral expression—has been a key issue for associative
learning theory from its inception. It is an obvious problem
because there are many cases in which behavior in the presence
of a stimulus belies what the animal knows about that stimulus.
Extinction, as noted above, is perhaps the most well-known
example, with behavior being eliminated after extinction, but the
A TRIBUTE TO ROBERT A. RESCORLA 253
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
original association can be unmasked through any number of
approaches (e.g., Bouton & Bolles, 1979;Delamater, 1996;
Rescorla, 1996a). Performance is a function of many factors and
there are many other demonstrations that performance itself is
not always a good readout of the state of the actual association
(e.g., Stout & Miller, 2007). For Bob, the learning-performance
issue was both a theoretical and a methodological one. The theo-
retical question was simply: how does learning map into per-
formance? For this question, Bob was generally agnostic and it
was never a question that he tried to answer directly. The meth-
odological question was, how can we make inferences that dif-
ferences in behavior truly reflect differences in learning,
however that mapping may happen? For this question, Bob
developed new approaches for evaluating performance experi-
mentally that led to many important discoveries about the nature
of basic learning processes.
For Bob, the problem of performance really boiled down to
whether we as a field were assessing learning under the appropri-
ate conditions to make reasonable inferences about learning from
ordinal differences in performance. He and his students identi-
fied several areas in which expression of learning in performance
is particularly dependent on how experimental questions are
asked. These include the observations that different CSs evoke
different CRs, that representations of the CS and US themselves
can alter expression of an association, that not knowing how
learning maps into performance places limitations on the infer-
ences that one can make from comparisons across different parts
of the behavioral scale, and that performance of an instrumental
response can be motivated by the Pavlovian cues present in a
given situation.
The Qualities of the CS Can Determine the Form
of the CR
The importance of understanding conditions for performance is
nicely illustrated in the dissertation work of some of Rescorla’s
students. Holland (1977) found that auditory and visual CSs that
are paired with delivery of a reinforcer result in different CRs
emerging during the CSs; auditory stimuli evoke head-jerking
whereas visual stimuli evoke orienting/rearing responses. An ex-
amination of a single performance measure (e.g., head jerking)
would result in an incorrect inference being made, for example,
that visual CSs evoke fewer CRs compared with auditory CSs.
Findings like these occurred in other preparations in Rescorla’s
lab, such as Durlach’s (1983) dissertation that also included orient-
ing to visual stimuli and hopper entry to auditory stimuli when she
was signaling intertrial USs in a pigeon autoshaping approach. In
Nairne and Rescorla’s (1981) work on second-order conditioning
described in the previous section, they demonstrated that even
though first-order auditory CSs did not evoke key pecking in
autoshaping, second-order CSs that predicted auditory CS rein-
forcers came to evoke pecking, showing that the association
between the auditory CS and the US was there, but not evident in
the pecking behavior that is typically measured in that task. In
reviewing work on learning in the spinal cord, Grau et al. (2022)
point out that Rescorla noted that one function of conditioning
might be to promote biologically preexisting response patterns
specific to a CS, but this learning is still subject to the same gen-
eral principles of learning. Together, work showing that the CS
may determine the form of the CR showed quite clearly that the
inferences that we might make about learning depend on what we
are measuring during a CS and whether the response we are meas-
uring is compatible with the qualitative features of that CS.
Nonassociative Contributions to Performance
When we think about performance tests, we typically think of
them as revealing the status of an association, but as noted above,
the measured response has to be appropriate for the CS and, as Bob
and his trainees showed in many studies, nonassociative changes in
the CS and US representations also may alter expression of that
association. For example, Rescorla (1973a) found that habituation
to the US would weaken conditioned responding (deflation effects)
and Rescorla (1974) found that exposure to a stronger US than was
used for conditioning would strengthen conditioned responding
(inflation effects). Rescorla and Skucy (1969) found that noncontin-
gent USs delivered during extinction maintained responding.
Rescorla and Heth (1975) showed that postextinction exposure to
the US from conditioning would reinstate extinguished conditioned
responding and Rescorla and Cunningham (1977) found that this
reinstatement effect could be weakened by extinction of a different
CS after US reexposure. Other research from that time by Rescorla
and Cunningham (1978a) found that the US representation is
depressed and then recovers over time in extinction. Later work by
Robbins (1990) showed that changes in attention to a CS may drive
extinction (loss of attention) and cause spontaneous recovery
(return of attention). In this special issue, Civile et al. (2022) evalu-
ate stimulus processing mechanisms in the context of human per-
ceptual learning and decision making. These findings together
show that a complete picture of the conditions in which associative
learning is expressed in performance needs to consider not just the
status of the CS-US association but must also include the potential
contribution of the conditioned response and nonassociative factors
within CS and US representations themselves.
The issue of what are the proper conditions to make inferences
about learning from performance was at the heart of all of Rescor-
la’s experiments. Most of his papers included replication of a key
effect in multiple experiments, with the first experiment using a
simple design to document the finding of interest, followed by
experiments including additional variables testing theoretical inter-
pretations. The experiments in his papers often replicated effects
with stimuli from different modalities, in Pavlovian and instru-
mental preparations, using between- and within-subjects designs,
and using different approaches in different species.
Disentangling Conditions for Learning From Conditions
for Performance
These approaches were always thorough and rigorous, but what
truly made Rescorla’s approach significant was that it always was
guided by a simple and elegant logic about how to conduct experi-
ments on animal learning. This approach involved giving an orga-
nism an experience at time 1 (t1) and then testing the impact of
that experience at time 2 (t2). In the most basic design, one group
would receive some sort of conditioning treatment while another
group would receive some sort of control treatment at t1, then both
groups would receive a test under common conditions at t2. Any
difference that occurs during that common test must be due to
254 COLWILL, DELAMATER, AND LATTAL
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
differential treatment at t1. The logic of this approach was outlined
in two influential reviews (Rescorla, 1988a;Rescorla & Holland,
1976). In Rescorla’s courses, he talked in great detail about a se-
ries of experiments from Davis and Wagner (1968,1969;Davis,
1970) that perfectly illustrate the need for common testing. Davis
and Wagner argued that much of the thinking about habituation at
the time was wrong because habituation experiments routinely did
not include a common test; comparisons between two groups
receiving different treatments were made during the course of ac-
quisition, when there was a confound between the conditions for
learning and performance. At that time, the thinking about habitua-
tion was that weaker stimuli produced greater habituation. But as
Davis and Wagner pointed out, these conclusions came from anal-
yses of acquisition curves that did not involve a common test—
thus, groups learned and were assessed under different conditions
for both learning and performance. For example, repeated presen-
tations of a 120 dB noise results in a high startle response that
decreases but continues to be higher than the startle response
evoked by a 108 dB noise, which starts off lower and ends at an
even lower point, leading to the conclusion that habituation is bet-
ter with weaker stimuli. When Davis and Wagner (1968) tested
the 120 dB and the 108 dB groups across a range of intensities in a
common test, the conclusions were reversed—the group habitu-
ated to 120 dB showed better habituation, revealed as a lower
response across a range of common intensities. Confounding the
conditions for learning with those for performance, which is an in-
herent problem in any study of acquisition curves, will lead to
inappropriate comparisons that may result in the wrong inference
about learning.
This t1-t2 issue and the problem of how to measure learning in
performance came up frequently in Rescorla’s work on studies of
acquisition and extinction. Rescorla had a long-standing interest in
the mechanisms of extinction—what caused extinction to develop
and what caused the behavior that was suppressed during extinc-
tion to return? In the late 1990s and early 2000s, Rescorla pub-
lished a series of studies evaluating mechanisms of spontaneous
recovery—the return of extinguished behavior following the pas-
sage of time. Much of this work was focused on theoretical mecha-
nisms, but Rescorla also wanted to change the framing of the way
we ask questions about spontaneous recovery experimentally.
In a large literature going back to Pavlov (1927), spontaneous
recovery experiments typically involved comparisons between two
different points in an animal’s life—the same type of comparisons
that plagued habituation experiments. Most frequently, spontane-
ous recovery is inferred from comparisons between behavior at the
end of extinction (t1) and the beginning of a test session (t2), with
the general finding that behavior increases at test. The problem
with this sort of approach as noted by Robbins (1990) in his disser-
tation in Rescorla’s lab is that the local conditions for performance
are different at the end of an extinction session and at the begin-
ning of a test. Another way to infer recovery is to give the orga-
nism two tests—one soon after extinction and one long after
extinction, with spontaneous recovery occurring if behavior is
greater during the second test compared with the first. Rescorla
(2004a) noted, however, that this sort of approach also has prob-
lems, “the tests of the two stimuli differ not only in the time since
extinction but also in the time since original training and in the
overall test context and age of the animal.”In many studies, these
two tests also involve repeated testing of the same stimulus,
meaning the first test functions as an additional extinction session
that may impact performance on the second test. Thus, the prob-
lem of comparing two test sessions soon and long after extinction
is problematic because the conditions for testing cannot possibly
be matched and include changes that may occur in the context that
are outside of experimental control. This sort of thinking led Bou-
ton (1991,1993) to propose that testing across different points in
time may result in testing in two very different contexts, resulting
in spontaneous recovery being a case of contextual renewal.
Rescorla solved this testing problem by using a within-subjects
design that holds the test day constant while manipulating the time
since extinction before that test (e.g., Rescorla, 1996b,1997b,
1997c,2005) and, in some cases, the time between acquisition and
extinction before the test (Rescorla, 2004b). This meant that he
could observe stimulus-specific spontaneous recovery in the same
animal in the same test session under identical conditions for per-
formance, which “controls for any general changes in the animal
or overall responding with the passage of time”(Rescorla, 1997b).
Although this design was a simple and elegant solution to the com-
mon testing problem in spontaneous recovery experiments, it
meant that learning and extinction would occur at different time
points, in contrast to most typical spontaneous recovery experi-
ments, in which acquisition and extinction occur at the same time
(Rescorla, 2004b). Time and retention are difficult variables and
there will always have to be differences in when a specific phase
of learning occurs. As noted in the Conditions section on appropri-
ate controls for Pavlovian conditioning, Rescorla thought that the
design of the experiment should be structured around the key
question of interest. For his work on spontaneous recovery, his in-
terest was in comparing the return of behavior with time, so it was
critical to match the conditions for performance at test and, ideally,
the result would hold using this approach and more traditional
approaches for measuring spontaneous recovery. In the present
special issue, Norton and Harris (2022) use some of the methods
advocated by Rescorla to examine the role of spontaneous recov-
ery type mechanisms in the partial reinforcement extinction effect.
The Scaling Problem in Assessing Performance
In designing spontaneous recovery experiments, Rescorla devel-
oped a clever way to test for spontaneous recovery in a single test
session. Other experimental questions are not as straightforward to
answer. In the 1990s and 2000s, Rescorla developed a procedure
to test a fundamental assumption about learning—that negatively
accelerated learning curves underlie conditioning and extinction.
Many studies have found that performance during acquisition and
extinction follows a negatively accelerated function and many the-
ories of learning, such as the Rescorla-Wagner model, predict that
the learning that underlies that performance also is negatively
accelerated. But Rescorla as an experimentalist argued that acqui-
sition and extinction curves themselves were not informative—
what we see in behavior does not necessarily reflect the underlying
learning. Indeed, in this issue Harris (2022) suggests some caveats
to consider in evaluating the shape of learning curves.
Theories like the Rescorla-Wagner model say that learning
curves are negatively accelerated because there should be a large
amount of learning early, when the discrepancy between the
expected and obtained US is large, followed by progressively
smaller changes as this discrepancy gets smaller and learning
A TRIBUTE TO ROBERT A. RESCORLA 255
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
approaches asymptote. The question is simple—is the change
from trial 1 to trial 2 greater than the change later in conditioning,
for example from trial 9 to trial 10? Simply comparing a change
from trial 1 to trial 2 with the change from trial 9 to trial 10 is
problematic because it involves a comparison against different
baselines (with trial 1 being lower and trial 9 being higher in per-
formance). There are statistical solutions for this type of problem
(attempting different ways of “normalizing”the data), but those
solutions simply hide the psychological problem of determining
what responding at different points in the scale means and how to
interpret deviations from those points in a sensible way (Stafford
& Lattal, 2009). This sort of issue and the different conclusions
that one may make based on transformations against different
baselines was illustrated in dissertations by Lattal (1999) and Got-
tlieb (2004) in Rescorla’s lab. As they noted, the only real solution
to this performance problem is to assess animals under common
conditions that equate baseline performance. In the cases of Lattal
and Gottlieb, this common testing was in conditions that attempted
to equate pre-CS responding in groups treated with different ITIs
or probabilities of reinforcement.
Rescorla’s solution to the problem of evaluating the negatively
accelerated learning curve assumption applied his common testing
logic to this issue by testing stimuli at different stages of condi-
tioning (early, when the associative change should be large or late,
when the change should be small). Rescorla (2001a) conducted a
series of very simple experiments that, for the first time, allowed
inferences to be made about the change in associative strength
early compared with late in learning. The key elements of the
design consisted of conditioning two stimuli separately for 10 ses-
sions (Aþand Cþ) in Phase 1, followed by one day of nonrein-
forced exposure to B and D (Band D) in Phase 2. At this
point, there would be no basis to expect any differences between
the AD and BC compounds. Thus, any difference that occurred
between those compounds after Phase 2 would have to be due to
the subsequent treatment in Phase 3, which consisted of three addi-
tional sessions of conditioning of Aþand, simultaneously, three
initial conditioning sessions of Bþ. By the end of acquisition, A
had received extensive conditioning (13 sessions) and B had
received less conditioning (three sessions). The Rescorla-Wagner
model predicts that at this point A should have more associative
strength than B, but that B should be undergoing more associative
change during those three sessions in Phase 3. To evaluate this,
Rescorla tested A and B in compound with D and C (examining
responding to the compound AD compared with BC). Prior to this
test, the total number of reinforced trials in the compound were the
same for AD and BC (13 sessions of A; 10 sessions of C and three
sessions of B). But the impact of those reinforcements should have
been higher in Phase 3 for B compared with A, resulting in greater
associative change in B and greater responding on the BC com-
pound compared with the AD compound at test. This is what he
found. He then adapted this procedure to apply the same logic to
testing whether greater changes in associative strength occur early
compared with late in extinction (Rescorla, 2001a).
As noted earlier, Rescorla used this compound test procedure to
test many fundamental assumptions about the nature of learning
(e.g., Rescorla, 2002b). Some of these assumptions (the shape of
the learning curve, the form of generalization gradients, how ele-
ments change in strength when conditioned in compound) have
been around since the dawn of learning theory, but until the early
2000s, had not been tested directly because of the inherent prob-
lem of comparing performance at different parts of the behavioral
scale. By testing compounds that were matched for performance to
the elements and overall history of those elements, Rescorla effec-
tively created common testing conditions in which behavioral dif-
ferences could not be easily explained away as differences in
performance. These compounds not only matched the local condi-
tions for performance, but they also matched the overall histories
of reinforcement and nonreinforcement of the elements in the
compound.
Pavlovian and Instrumental Discriminative Stimulus
Effects on Expression of Instrumental Responding
Although much of Bob’s work focused on understanding how
Pavlovian learning is expressed in Pavlovian conditioned responses,
he had a long-standing interest in understanding how Pavlovian
CSs influenced performance of instrumental behavior, a finding
first documented by Estes and Skinner (1941).Bob’s two-process
theory published in 1967 with Solomon was a beautiful piece of
scholarship that reviewed theorizing up to that point on the ways in
which Pavlovian CSs influence instrumental behavior (Rescorla &
Solomon, 1967). That article included a review of recent work
from the Solomon laboratory from Overmier, Lolordo, and
others showing the impact of Pavlovian stimuli on instrumental
responding and led to the general idea that instrumental respond-
ing can be motivated by Pavlovian CSs that are present when the
opportunity to respond is presented. This led to the idea of Pav-
lovian-to-instrumental transfer (PIT), which has been used in the
years since to delineate the content of associative learning, as
reviewed above.
Many experiments from Bob’s lab demonstrated that not only
do Pavlovian CSs augment instrumental responding, but they do
so more when the Pavlovian CSs and instrumental responses lead
to the same outcome, as reviewed above. This outcome-specific
transfer effect has been critical in delineating the associative con-
tent of learning. But as with many of Bob’s experiments, he cre-
ated a set of tools that can be applied to answering different
questions about the expression of learning in performance. For
example, many studies have found various ways to arrange extinc-
tion (e.g., nonreinforcement, contingency degradation) that can
eliminate conditioned responding, but few studies have been able
to probe the content of that original association while conditioned
responding is still extinguished. Most experiments need some sort
of unmasking procedure, such as reinstatement, the passage of
time, or a change in context to reveal the association through the
return of the Pavlovian behavior that was extinguished. In a series
of experiments conducted in Rescorla’s lab, Delamater (1996)
found that extinction of Pavlovian CSs had no effect on the ability
of those CSs to transfer to instrumental responses that shared the
same outcomes, even when the Pavlovian CRs remained extin-
guished. This showed that animals maintained that Pavlovian asso-
ciation even when it was not evident in behavior as a Pavlovian
CR, demonstrating a clear distinction between what the animal
had learned (as revealed in a specific PIT response) and what the
animal expressed in behavior (as revealed in extinguished Pavlov-
ian responses). As Rescorla (2003a) noted, findings like these go a
long way toward demonstrating what Pavlov (1927) thought, but
256 COLWILL, DELAMATER, AND LATTAL
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
never showed—that extinction leaves the original learning com-
pletely intact.
Rescorla used specific transfer techniques in other creative
ways to show how instrumental discriminative stimuli may influ-
ence instrumental performance, demonstrating, for example, that
presenting two discriminative stimuli in compound produces more
summation when each signals the same R-O relations compared
with when they signal different R-O relations (Rescorla, 1990a).
Findings that transfer does not depend only on shared outcomes
between stimuli and responses, but also on shared R-O relations is
consistent with a hierarchical framework for thinking about
instrumental performance. Approaches like this illustrate that
Pavlovian conditioning is important not only in its own right, but
gives us behavioral approaches for exploring other forms of
learning (Rescorla, 1987b).
As with everything that Bob did, his approach to how learning
is expressed in performance left us with many fundamental find-
ings and a set of tools that we can use to ask questions about learn-
ing that often seem intractable. Although he never answered the
question of how learning is specifically mapped in performance
(nor did he really ever try to), he made a clear case through his rig-
orous approach that inferences about learning can be made from
performance when we ask the experimental questions in the right
way. He left us with associative and nonassociative factors to con-
sider, as well as a set of tools that we can use to make reasonable
inferences that differences in behavior reflect differences in
learning.
Final Thoughts
We have not attempted to summarize all of Rescorla’s work or
contributions, which would be impossible in this introduction to
the special issue. We have tried instead to provide an overview of
Rescorla’s influence on the field centered around the organiza-
tional structure that he believed was the proper framework for
thinking about learning. We hope that we have conveyed a good
sense of his astounding legacy of empirical and theoretical contri-
butions to learning over a career spanning five decades. Rescorla
opened our eyes to new ways of thinking about learning and how
to frame questions and design experiments to shed new light on
enduring problems as well as to point to new directions.
When Bob retired in 2009, he left behind emerging controver-
sies and important but unanswered questions. We have identified
four questions that we think he would have thought especially
worthwhile to pursue. First, there is no doubt that he would agree
that much remains to be discovered about the mapping of learning
into performance. Somewhat ironically, perhaps acknowledging
Hull’s struggles with the mapping of learning into performance,
Rescorla and Holland (1982) argued that theoretical developments
had at the time advanced in learning precisely because the ques-
tion of “how learning is exhibited in performance”had been post-
poned but they called for the issue to be addressed. Rescorla
(2001a) echoed that call and pointed out that “the absence of well-
specified rules for mapping learning into performance makes it
difficult to infer changes in the size of associative changes for
stimuli occupying different points on the performance scale”and
he devised the compound test to address the issue. He thought
more work was needed to understand even the most basic ques-
tions about summation rules for stimulus compounds and the
shape of the learning curve. McNally et al. (2011), for example,
represents a good step in the right direction as they grapple with
the competition between motivational systems and predicting
performance.
Second, he would surely have encouraged consideration of the
question about whether the same mechanism underlies all forms of
inhibitory learning. In one of his last articles, Rescorla reported an
extinction-like renewal effect with overexpectation. He thought
that commonality between the decrements produced by nonrein-
forcement and overexpectation “could have important implications
for the description of the conditions producing extinction”
(Rescorla, 2007). However, recent research by Lay et al. (2019)
pointing to a dissociation in the underlying mechanisms of extinc-
tion and overexpectation is the sort of research that begins to
address this important question. Further, he noted that his results
did not “shed light on the nature of the decremental mechanism
itself”(Rescorla, 2007), and he left us with several different types
of mechanisms to account for extinction phenomena. These
include nonassociative changes in stimuli and outcomes, true asso-
ciative weakening, new inhibitory learning (e.g., inhibitory S-R
associations), new excitatory learning about the US’s motivation-
ally opposite state (e.g., in the case of appetitive learning the S or
R could associate with a state of “frustration”), and new hierarchi-
cal learning where the stimulus acquires control over a new associ-
ation between the R and “frustration.”His concern about our
primitive understanding of extinction was a recurrent theme in his
writing and he would most certainly applaud efforts to understand
the nature of the learning that occurs about the responses made in
extinction (Rescorla, 2001b).
Third, there is a real need to understand how the various binary
and hierarchical associations in instrumental learning interact and
contribute to performance. Although earlier work in his lab clearly
demonstrated the existence of a wide variety of binary and hier-
archical structures that mediate instrumental learning, the difficult
work of establishing a clear theoretical framework for thinking
about their interactions remains. Bob made some progress on this
problem looking at the interaction of instrumental stimuli and
responses with multiple outcomes (Rescorla, 1991a,1992a). With
regard to S-O and R-O associations, he concluded that “there is
surprisingly little interference among the associations developed
in instrumental learning”(Rescorla, 1992a). Much of the modern
work on this problem, for instance, emphasizes control by S-R and
R-O mechanisms (e.g., Perez & Dickinson, 2020), but our view is
that this strict focus on what has come to be called “actions”and
“habits”does not begin to address the sorts of complexities
revealed by Rescorla and his colleagues’work on the associative
structures that underlies instrumental learning. Clearly, this is an
avenue ripe for development.
Fourth, he would be pleased to see attention focused on the
global versus individual error term debate. Rescorla (2003a) noted
the ubiquity of the basic idea of prediction error in modern neural
network approaches and that most error correction models assume
that the same error applies to all stimuli treated together, “indeed,
it is that feature that makes neural network models so powerful.”
He thought that if this assumption could continue to be challenged
that it could potentially be “disastrous for such models.”Bob was
especially interested in understanding if there were conditions in
which stimuli treated in compound could change in opposite direc-
tions. If such a finding were demonstrated, it would cause a
A TRIBUTE TO ROBERT A. RESCORLA 257
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
dramatic shift in our thinking about the ways in which prediction
error works. His own research, of course, revealed differences in
the amount of associative gain or loss between stimuli conditioned
in compound as a function of their individual prediction discrepan-
cies. Whether or not this would require a wholesale overhaul of
learning and computational models (e.g., Holmes et al., 2019; see
in this issue Honey et al., 2022) remains an important topic.
Bob was keenly aware that the field would keep moving for-
ward without him, and when he retired in 2009 he was extremely
humble about how lasting his contributions would be. In our view,
this was one of the few wrong conclusions that Bob made—he
will forever be mentioned in the same breath as Pavlov and other
luminaries of the field. But Bob recognized that his students and
other students of learning theory would be the ones to generate the
new knowledge that would expand on what he has given us.
Knowing this, Rescorla was extraordinarily generous with his time
and spent a lot of it in the lab talking with his students and col-
leagues about learning. What he especially appreciated were con-
versations that moved a question forward, regardless of whether
they spanned hours, days, or weeks. He wanted his students to
keep thinking about a question and come back with something
new to say to which he would always have lightning-like
responses, most of which would require a period of silent intro-
spection to prepare a response. Although Bob often had the last
word, he really hoped that his work would inspire the way forward
for others, as it surely has.
References
Adams, C. D., & Dickinson, A. (1981). Actions and habits: Variations in
associative representations during instrumental learning. In N. E. Spear
& R. R. Miller (Eds.), Information processing in animals: Memory
mechanisms (pp. 143–165). Erlbaum.
Aggarwal, M., Akamine, Y., Liu, A. W., & Wickens, J. R. (2020). The nu-
cleus accumbens and inhibition in the ventral tegmental area play a
causal role in the Kamin blocking effect. European Journal of Neuro-
science,52(3), 3087–3109. https://doi.org/10.1111/ejn.14732
Alarcón, D. E., & Delamater, A. R. (2019). Outcome-specific Pavlovian-
to-instrumental transfer (PIT) with alcohol cues and its extinction. Alco-
hol,76, 131–146. https://doi.org/10.1016/j.alcohol.2018.09.003
Asratyan, E. A. (1965). Compensatory adaptation, reflex activity, and the
brain (S. A. Corson, Trans.). Oxford University Press.
Austen, J. M., Sprengel, R., & Sanderson, D. J. (2022). Reinforcement rate
and the balance between excitatory and inhibitory learning: Insights
from deletion of the GluA1 AMPA receptor subunit. Journal of Experi-
mental Psychology: Animal Learning & Cognition,48(4), 307–314.
https://doi.org/10.1037/xan0000336
Balleine, B. W. (2019). The meaning of behavior: Discriminating reflex
and volition in the brain. Neuron,104(1), 47–62. https://doi.org/10
.1016/j.neuron.2019.09.024
Barnet, R. C., Arnold, H. M., & Miller, R. R. (1991). Simultaneous condi-
tioning demonstrated in second-order conditioning: Evidence for similar
associative structure in forward and simultaneous conditioning. Learn-
ing and Motivation,22(3), 253–268. https://doi.org/10.1016/0023-
9690(91)90008-V
Bengtsson, F., Jirenhed, D.-A., Svensson, P., & Hesslow, G. (2007).
Extinction of conditioned blink responses by cerebello-olivary pathway
stimulation. Neuroreport,18(14), 1479–1482. https://doi.org/10.1097/
WNR.0b013e3282e326e8
Bengtsson, F., Svensson, P., & Hesslow, G. (2004). Feedback control of
Purkinje cell activity by the cerebello-olivary pathway. The European
Journal of Neuroscience,20(11), 2999–3005. https://doi.org/10.1111/j
.1460-9568.2004.03789.x
Bolles, R. C. (1972). Reinforcement, expectancy, and learning. Psycholog-
ical Review,79(5), 394–409. https://doi.org/10.1037/h0033120
Bouton, M. E. (1991). Context and retrieval in extinction and in other exam-
ples of interference in simple associative learning. In L. Dachowski &
C. F. Flaherty (Eds.), Current topics in animal learning: Brain, emotion,
and cognition (pp. 25–53). Erlbaum.
Bouton, M. E. (1993). Context, time, and memory retrieval in the interfer-
ence paradigms of Pavlovian learning. Psychological Bulletin,114(1),
80–99. https://doi.org/10.1037/0033-2909.114.1.80
Bouton, M. E., & Bolles, R. C. (1979). Contextual control of the extinction
of conditioned fear. Learning and Motivation,10(4), 445–466. https://
doi.org/10.1016/0023-9690(79)90057-2
Brandon, S. E., Vogel, E. H., & Wagner, A. R. (2000). A componential
view of configural cues in generalization and discrimination in Pavlov-
ian conditioning. Behavioural Brain Research,110(1–2), 67–72. https://
doi.org/10.1016/S0166-4328(99)00185-0
Bush, R. R., & Mosteller, F. (1951). A mathematical model for simple
learning. Psychological Review,58(5), 313–323. https://doi.org/10
.1037/h0054388
Cartoni, E., Balleine, B., & Baldassarre, G. (2016). Appetitive Pavlovian-
instrumental transfer: A review. Neuroscience and Biobehavioral
Reviews,71, 829–848. https://doi.org/10.1016/j.neubiorev.2016.09.020
Chan, Y. Y., Westbrook, R. F., & Holmes, N. M. (2021). Protecting the
Rescorla-Wagner (1972) theory: A reply to Spicer et al. (2020). Journal
of Experimental Psychology: Animal Learning and Cognition,47(2),
211–215. https://doi.org/10.1037/xan0000271
Chen, J. S., & Amsel, A. (1980). Recall (versus recognition) of taste and
immunization against aversive taste anticipations based on illness. Sci-
ence,209(4458), 831–833. https://doi.org/10.1126/science.7403850
Civile, C., McLaren, R., Cooke, A., & McLaren, I. P. L. (2022). Using
transcranial direct current stimulation (tDCS) to influence decision crite-
rion in a target detection paradigm. Journal of Experimental Psychol-
ogy: Animal Learning & Cognition,48(4), 383–395. https://doi.org/10
.1037/xan0000329
Colwill, R. M. (1991). Negative discriminative stimuli provide information
about the identity of omitted response-contingent outcomes. Animal Learn-
ing & Behavior,19(4), 326–336. https://doi.org/10.3758/BF03197893
Colwill, R. M. (1994). Associative representations of instrumental contin-
gencies. In D. L. Medin (Ed.), The psychology of learning and motiva-
tion (pp. 1–72). Academic Press.
Colwill, R. M. (2007). Effect of the passage of time on the contribution of
initial response-outcome associations to instrumental performance.
Behavioural Processes,75(1), 14–22. https://doi.org/10.1016/j.beproc
.2007.01.008
Colwill, R. M. (2019). Behavioral studies of stimulus learning in zebrafish
larvae. Behavioural Processes,164, 150–156.
Colwill, R. M., & Delamater, B. D. (1995). An associative analysis of
instrumental biconditional discrimination learning. Animal Learning &
Behavior,23(2), 218–233. https://doi.org/10.3758/BF03199937
Colwill, R. M., & Rescorla, R. A. (1985a). Extensive training, partial rein-
forcement, and temporal gaps do not affect S-S learning in second-order
autoshaping. Animal Learning & Behavior,13(2), 165–170. https://doi
.org/10.3758/BF03199269
Colwill, R. M., & Rescorla, R. A. (1985b). Instrumental responding
remains sensitive to reinforcer devaluation after extensive training. Jour-
nal of Experimental Psychology: Animal Behavior Processes,11(4),
520–536. https://doi.org/10.1037/0097-7403.11.4.520
Colwill, R. M., & Rescorla, R. A. (1985c). Post-conditioning devaluation
of a reinforcer affects instrumental responding. Journal of Experimental
Psychology: Animal Behavior Processes,11(1), 120–132. https://doi
.org/10.1037/0097-7403.11.1.120
258 COLWILL, DELAMATER, AND LATTAL
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
Colwill, R. M., & Rescorla, R. A. (1986). Associative structures in instru-
mental learning. In G. H. Bower (Ed.), The psychology of learning and
motivation (Vol. 20, pp. 55–104). Elsevier.
Colwill, R. M., & Rescorla, R. A. (1988a). Associations between the dis-
criminative stimulus and the reinforcer in instrumental learning. Journal
of Experimental Psychology: Animal Behavior Processes,14(2), 155–
164. https://doi.org/10.1037/0097-7403.14.2.155
Colwill, R. M., & Rescorla, R. A. (1988b). The role of response-reinforcer
associations increases throughout extended instrumental training. Animal
Learning & Behavior,16(1), 105–111. https://doi.org/10.3758/BF03209051
Colwill, R. M., & Rescorla, R. A. (1990a). Effect of reinforcer devaluation
on discriminative control of instrumental behavior. Journal of Experi-
mental Psychology: Animal Behavior Processes,16(1), 40–47. https://
doi.org/10.1037/0097-7403.16.1.40
Colwill, R. M., & Rescorla, R. A. (1990b). Evidence for the hierarchical
structure of instrumental learning. Animal Learning & Behavior,18(1),
71–82. https://doi.org/10.3758/BF03205241
Colwill, R. M., & Triola, S. M. (2002). Instrumental responding remains
under the control of the consequent outcome after extended training.
Behavioural Processes,57(1), 51–64. https://doi.org/10.1016/S0376-
6357(01)00204-2
Cooper, L. D., Aronson, L., Balsam, P. D., & Gibbon, J. (1990). Duration
of signals for intertrial reinforcement and nonreinforcement in random
control procedures. Journal of Experimental Psychology: Animal Behav-
ior Processes,16(1), 14–26. https://doi.org/10.1037/0097-7403.16.1.14
Corbit, L. H., & Balleine, B. W. (2005). Double dissociation of basolateral
and central amygdala lesions on the general and outcome-specific forms
of pavlovian- instrumental transfer. Journal of Neuroscience,25(4),
962–970. https://doi.org/10.1523/JNEUROSCI.4507-04.2005
Corbit, L. H., & Balleine, B. W. (2011). The general and outcome-specific
forms of Pavlovian-instrumental transfer are differentially mediated by
the nucleus accumbens core and shell. Journal of Neuroscience,31(33),
11786–11794. https://doi.org/10.1523/JNEUROSCI.2711-11.2011
Coutureau, E., & Killcross, S. (2003). Inactivation of the infralimbic pre-
frontal cortex reinstates goal-directed responding in overtrained rats.
Behavioural Brain Research,146(1–2), 167–174. https://doi.org/10.1016/
j.bbr.2003.09.025
Crimmins, B., Burton, T. J., McNulty, M., Laurent, V., Hart, G., &
Balleine, B. W. (2022). Response-independent outcome presentations
weaken the instrumental response-outcome association. Journal of Ex-
perimental Psychology: Animal Learning & Cognition,48(4), 396–412.
https://doi.org/10.1037/xan0000340
Davidson, T. L., Aparicio, J., & Rescorla, R. A. (1988). Transfer between Pav-
lovian facilitators and instrumental discriminative stimuli. Animal Learning
&Behavior,16(3), 285–291. https://doi.org/10.3758/BF03209078
Davidson, T. L., & Rescorla, R. A. (1986). Transfer of facilitation in the
rat. Animal Learning & Behavior,14(4), 380–386. https://doi.org/10
.3758/BF03200082
Davis, M. (1970). Effects of interstimulus interval length and variability on
startle-response habituation in the rat. Journal of Comparative and Physio-
logical Psychology,72(2), 177–192. https://doi.org/10.1037/h0029472
Davis, M., & Wagner, A. R. (1968). Startle responsiveness after habitua-
tion to different intensities of tone. Psychonomic Science,12(7), 337–
338. https://doi.org/10.3758/BF03331339
Davis, M., & Wagner, A. R. (1969). Habituation of startle response under
incremental sequence of stimulus intensities. Journal of Comparative
and Physiological Psychology,67(4), 486–492. https://doi.org/10.1037/
h0027308
Delamater, A. R. (1996). Effects of several extinction treatments upon the
integrity of Pavlovian stimulus-outcome associations. Animal Learning
& Behavior,24(4), 437–449. https://doi.org/10.3758/BF03199015
Delamater, A. R., & Lattal, K. M. (2014). The study of associative learning:
Mapping from psychological to neural levels of analysis. Neurobiology of
Learning and Memory,108,1–4. https://doi.org/10.1016/j.nlm.2013.12
.006
Delamater, A. R., Schneider, K., & Derman, R. C. (2017). Extinction of
specific stimulus-outcome (S-O) associations in Pavlovian learning with
an extended CS procedure. Journal of Experimental Psychology: Ani-
mal Learning and Cognition,43(3), 243–261. https://doi.org/10.1037/
xan0000138
Delamater, A. R., & Westbrook, R. F. (2014). Psychological and neural
mechanisms of experimental extinction: A selective review. Neurobiology
of Learning and Memory,108,38–51. https://doi.org/10.1016/j.nlm.2013
.09.016
Dickinson, A. (1980). Contemporary animal learning theory. Cambridge
University Press.
Dickinson, A., & Charnock, D. J. (1985). Contingency effects with main-
tained instrumental reinforcement. The Quarterly Journal of Experimen-
tal Psychology: B, Comparative and Physiological Psychology,
37B(4b), 397–416. https://doi.org/10.1080/14640748508401177
Dickinson, A., Hall, G., & Mackintosh, N. J. (1976). Surprise and the
attenuation of blocking. Journal of Experimental Psychology: Animal
Behavior Processes,2(4), 313–322. https://doi.org/10.1037/0097-7403.2
.4.313
Domjan, M. (2005). Pavlovian conditioning: A functional perspective. An-
nual Review of Psychology,56(1), 179–206. https://doi.org/10.1146/
annurev.psych.55.090902.141409
Domjan, M. (2015). The Garcia-Koelling selective association effect: A
historical and personal perspective. International Journal of Compara-
tive Psychology,28, 25645. https://doi.org/10.46867/ijcp.2015.28.01.08
Durlach, P. J. (1983). Effect of signaling intertrial unconditioned stimuli in
autoshaping. Journal of Experimental Psychology: Animal Behavior
Processes,9(4), 374–389. https://doi.org/10.1037/0097-7403.9.4.374
Durlach, P. J., & Rescorla, R. A. (1980). Potentiation rather than overshadow-
ing in flavor-aversion learning: An analysis in terms of within-compound
associations. Journal of Experimental Psychology: Animal Behavior Proc-
esses,6(2), 175–187. https://doi.org/10.1037/0097-7403.6.2.175
Estes, W. K., & Skinner, B. F. (1941). Some quantitative properties of anx-
iety. Journal of Experimental Psychology,29(5), 390–400. https://doi
.org/10.1037/h0062283
Fam, J., Chieng, B., Westbrook, R. F., Laurent, V., & Holmes, N. M.
(2022). Second-order fear conditioning involves formation of competing
stimulus-danger and stimulus-safety associations. Cerebral Cortex.
Advance online publication. https://doi.org/10.1093/cercor/bhac176
Fanselow, M. S. (1998). Pavlovian conditioning, negative feedback, and
blocking: Mechanisms that regulate association formation. Neuron,
20(4), 625–627. https://doi.org/10.1016/S0896-6273(00)81002-8
Gallistel, C. R., & Gibbon, J. (2000). Time, rate, and conditioning. Psycho-
logical Review,107(2), 289–344. https://doi.org/10.1037/0033-295X
.107.2.289
Garcia, J., & Koelling, R. A. (1966). Relation of cue to consequence in
avoidance learning. Psychonomic Science,4(1), 123–124. https://doi
.org/10.3758/BF03342209
Garr, E., Bushra, B., Tu, N., & Delamater, A. R. (2020). Goal-directed
control on interval schedules does not depend on the action-outcome
correlation. Journal of Experimental Psychology: Animal Learning and
Cognition,46(1), 47–64. https://doi.org/10.1037/xan0000229
Garr, E., Padovan-Hernandez, Y., Janak, P. H., & Delamater, A. R.
(2021). Maintained goal-directed control with overtraining on ratio
schedules. Learning & Memory,28(12), 435–439. https://doi.org/10
.1101/lm.053472.121
Gibbon, J., & Balsam, P. (1981). Spreading association in time. In C. M.
Locurto, H. S. Terrace, J. Gibbon (Eds.), Autoshaping and conditioning
theory (pp. 219–253). Academic Press.
Gostolupce, D., Lay, B. P. P., Maes, E. J. P., & Iordanova, M. D. (2022).
Understanding associative learning through higher-order conditioning.
A TRIBUTE TO ROBERT A. RESCORLA 259
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
Frontiers in Behavioral Neuroscience,16, Article 845616. https://doi
.org/10.3389/fnbeh.2022.845616
Gottlieb, D. A. (2004). Acquisition with partial and continuous reinforce-
ment in pigeon autoshaping. Learning & Behavior,32(3), 321–334.
https://doi.org/10.3758/BF03196031
Grand, C., Close, J., Hale, J., & Honey, R. C. (2007). The role of similarity
in human associative learning. Journal of Experimental Psychology:
Animal Behavior Processes,33(1), 64–71. https://doi.org/10.1037/0097-
7403.33.1.64
Grau, J. W., Hudson, K. E., Tarbet, M. M., & Strain, M. M. (2022). Behav-
ioral studies of spinal conditioning The spinal cord is smarter than you
think it is. Journal of Experimental Psychology: Animal Learning &
Cognition,48(4), 435–457. https://doi.org/10.1037/xan0000332
Grau, J. W., & Rescorla, R. A. (1984). Role of context in autoshaping.
Journal of Experimental Psychology: Animal Behavior Processes,
10(3), 324–332. https://doi.org/10.1037/0097-7403.10.3.324
Gupta, T. A., Daniels, C. W., Espinoza, J. I., Smith, B. H., & Sanabria, F.
(2022). Assessing complex odor discrimination in mice using a novel
instrumental patterning task. Journal of Experimental Psychology: Ani-
mal Learning & Cognition,48(4), 358–369. https://doi.org/10.1037/
xan0000341
Haas, R. B., Shessel, F. M., Willner, H. S., & Rescorla, R. A. (1970). The
effect of satiation following partial reinforcement. Psychonomic Science,
18(5), 296–297. https://doi.org/10.3758/BF03331835
Hall, G. (2022). Extinction of conditioned flavor preferences. Journal of Ex-
perimental Psychology: Animal Learning & Cognition,48(4), 349–357.
https://doi.org/10.1037/xan0000326
Harris, J. A. (2022). The learning curve, revisited. Journal of Experimental
Psychology: Animal Learning & Cognition,48(4), 265–280. https://doi
.org/10.1037/xan0000331
Hesslow, G., & Ivarsson, M. (1996). Inhibition of the inferior olive during
conditioned responses in the decerebrate ferret. Experimental Brain
Research,110(1), 36–46. https://doi.org/10.1007/BF00241372
Higgins, T., & Rescorla, R. A. (2004). Extinction and retraining of simulta-
neous and successive flavor conditioning. Learning & Behavior,32(2),
213–219. https://doi.org/10.3758/BF03196022
Hogarth, L., Dickinson, A., Wright, A., Kouvaraki, M., & Duka, T. (2007).
The role of drug expectancy in the control of human drug seeking. Jour-
nal of Experimental Psychology: Animal Behavior Processes,33(4),
484–496. https://doi.org/10.1037/0097-7403.33.4.484
Holland, P. C. (1977). Conditioned stimulus as a determinant of the form
of the Pavlovian conditioned response. Journal of Experimental Psy-
chology: Animal Behavior Processes,3(1), 77–104. https://doi.org/10
.1037/0097-7403.3.1.77
Holland, P. C., & Rescorla, R. A. (1975a). The effects of two ways of
devaluing the unconditioned stimulus after first- and second-order con-
ditioning. Journal of Experimental Psychology: Animal Behavior Proc-
esses,88(4), 459–467. https://doi.org/10.1037/0097-7403.1.4.355
Holland, P. C., & Rescorla, R. A. (1975b). Second-order conditioning with
food unconditioned stimulus. Journal of Comparative and Physiological
Psychology,88(1), 459–467. https://doi.org/10.1037/h0076219
Holmes,N.M.,Cai,S.Y.,Lay,B.P.P.,Watts,N.R.,&Westbrook,R.F.
(2014). Extinguished second-order conditioned fear responses are renewed
but not reinstated. Journal of Experimental Psychology: Animal Learning
and Cognition,40(4), 440–456. https://doi.org/10.1037/xan0000036
Holmes, N. M., Chan, Y. Y., & Westbrook, R. F. (2019). A combination
of common and individual error terms is not needed to explain associa-
tive changes when cues with different training histories are conditioned
in compound: A review of Rescorla’s compound test procedure. Journal
of Experimental Psychology: Animal Learning and Cognition,45(2),
242–256. https://doi.org/10.1037/xan0000204
Holmes, N. M., Marchand, A. R., & Coutureau, E. (2010). Pavlovian to
instrumental transfer: A neurobehavioural perspective. Neuroscience
and Biobehavioral Reviews,34(8), 1277–1295. https://doi.org/10.1016/j
.neubiorev.2010.03.007
Honey, R. C., Dwyer, D. M., & Iliescu, A. F. (2022). Associative change
in Pavlovian conditioning: A reappraisal. Journal of Experimental Psy-
chology: Animal Learning & Cognition,48(4), 281–294. https://doi.org/
10.1037/xan0000318
Hull, C. L. (1943). Principles of behavior. Appleton-Century-Crofts.
Iordanova, M. D., Yau, J. O. Y., McDannald, M. A., & Corbit, L. H.
(2021). Neural substrates of appetitive and aversive prediction error.
Neuroscience and Biobehavioral Reviews,123, 337–351. https://doi.org/
10.1016/j.neubiorev.2020.10.029
Jones, P. M., Mitchell, C. J., Wills, A. J., & Spicer, S. G. (2021). Similar-
ities and differences: Comment on Chan et al. (2021). Journal of Experi-
mental Psychology: Animal Learning and Cognition,47(2), 216–217.
https://doi.org/10.1037/xan0000277
Kamin, L. J. (1968). “Attention-like”processes in classical conditioning.
In M. R. Jones (Ed.), Miami Symposium on the prediction of behavior:
Aversive stimulation (pp. 9–33). University of Miami Press.
Killcross, S., & Coutureau, E. (2003). Coordination of actions and habits
in the medial prefrontal cortex of rats. Cerebral Cortex,13(4), 400–408.
https://doi.org/10.1093/cercor/13.4.400
Kim, J. J., Krupa, D. J., & Thompson, R. F. (1998). Inhibitory cerebello-
olivary projections and blocking effect in classical conditioning. Sci-
ence,279(5350), 570–573. https://doi.org/10.1126/science.279.5350
.570
Kosaki, Y., & Dickinson, A. (2010). Choice and contingency in the devel-
opment of behavioral autonomy during instrumental conditioning. Jour-
nal of Experimental Psychology: Animal Behavior Processes,36(3),
334–342. https://doi.org/10.1037/a0016887
Krane, R. V., & Wagner, A. R. (1975). Taste aversion learning with a
delayed shock U.S.: Implications for the “Generality of the laws of
learning.”Journal of Comparative and Physiological Psychology,88(2),
882–889. https://doi.org/10.1037/h0076417
Lattal, K. M. (1999). Trial and intertrial durations in Pavlovian condition-
ing: Issues of learning and performance. Journal of Experimental Psy-
chology: Animal Behavior Processes,25(4), 433–450. https://doi.org/10
.1037/0097-7403.25.4.433
Lattal, K. M. (2013). Pavlovian conditioning. In G. J. Madden (Ed.), APA
handbook of behavior analysis: Vol. 1 Methods and principles (pp. 283–
306). American Psychological Association Press.
Lattal, K. M., & Nakajima, S. (1998). Overexpectation in appetitive Pav-
lovian and instrumental conditioning. Animal Learning & Behavior,
26(3), 351–360. https://doi.org/10.3758/BF03199227
Laurent, V., & Balleine, B. W. (2015). Factual and counterfactual action-
outcome mappings control choice between goal-directed actions in rats.
Current Biology,25(8), 1074–1079. https://doi.org/10.1016/j.cub.2015
.02.044
Laurent, V., Leung, B., Maidment, N., & Balleine, B. W. (2012).
m
- and
d
-opioid-related processes in the accumbens core and shell differentially
mediate the influence of reward-guided and stimulus-guided decisions
on choice. The Journal of Neuroscience,32(5), 1875–1883. https://doi
.org/10.1523/JNEUROSCI.4688-11.2012
Laurent, V., Priya, P., Crimmins, B. E., & Balleine, B. W. (2021). General
Pavlovian-instrumental transfer tests reveal selective inhibition of the
response type - whether Pavlovian or instrumental - performed during
extinction. Neurobiology of Learning and Memory,183, 107483. https://
doi.org/10.1016/j.nlm.2021.107483
Lay, B. P. P., Nicolosi, M., Usypchuk, A. A., Esber, G. R., & Iordanova,
M. D. (2019). Dissociation of appetitive overexpectation and extinction
in the infralimbic cortex. Cerebral Cortex,29(9), 3687–3701. https://doi
.org/10.1093/cercor/bhy248
Lichtenberg, N. T., Pennington, Z. T., Holley, S. M., Greenfield, V. Y.,
Cepeda, C., Levine, M. S., & Wassum, K. M. (2017). Basolateral amyg-
dala to orbitofrontal cortex projections enable cue-triggered reward
260 COLWILL, DELAMATER, AND LATTAL
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
expectations. The Journal of Neuroscience,37(35), 8374–8384. https://
doi.org/10.1523/JNEUROSCI.0486-17.2017
Lindblom, L. L., & Jenkins, H. M. (1981). Responses eliminated by non-
contingent or negatively contingent reinforcement recover in extinction.
Journal of Experimental Psychology: Animal Behavior Processes,7(2),
175–190. https://doi.org/10.1037/0097-7403.7.2.175
LoLordo, V. M. (1979). Selective associations. In A. Dickinson & R. A.
Boakes (Eds.), Mechanisms of learning and motivation: A memorial vol-
ume to Jerzy Konorski (pp. 367–398). Erlbaum.
LoLordo, V. M., & Droungas, A. (1989). Selective associations and adapt-
ive specializations: Taster aversions and phobias. In S. B. Klein & R. R.
Mowrer (Eds.), Contemporary learning theories: Instrumental condi-
tioning theory and the impact of biological constraints on learning (pp.
145–179). Erlbaum.
Lovibond, P. F., Chow, J. Y. L., Tobler, C., & Lee, J. C. (2022). Reversal
of inhibition by no-modulation training but not extinction in human
causal learning. Journal of Experimental Psychology: Animal Learning
& Cognition,48(4), 336–348. https://doi.org/10.1037/xan0000328
Mackintosh, N. J. (1975). A theory of attention: Variations in the associ-
ability of stimuli with reinforcement. Psychological Review,82(4), 276–
298. https://doi.org/10.1037/h0076778
Mackintosh, N. J. (1983). Conditioning and associative learning. Claren-
don Press.
Mackintosh, N. J., & Dickinson, A. (1979). Instrumental (Type II) condi-
tioning. In A. Dickinson & R. A. Boakes (Eds.), Mechanisms of learning
and motivation (pp. 143–169). Erlbaum.
Marchant, N., & Chaigneau, S. E. (2022). On the importance of feedback for
categorization: Revisiting category learning experiments using an Adaptive
Filter model. Journal of Experimental Psychology: Animal Learning &
Cognition,48(4), 295–306. https://doi.org/10.1037/xan0000339
McNally, G. P., Johansen, J. P., & Blair, H. T. (2011). Placing prediction
into the fear circuit. Trends in Neurosciences,34(6), 283–292. https://
doi.org/10.1016/j.tins.2011.03.005
Miller, R. R., Barnet, R. C., & Grahame, N. J. (1995). Assessment of the
Rescorla-Wagner model. Psychological Bulletin,117(3), 363–386.
https://doi.org/10.1037/0033-2909.117.3.363
Morgan, M. J. (1974). Resistance to satiation. Animal Behaviour,22(2),
449–466. https://doi.org/10.1016/S0003-3472(74)80044-8
Morgan, M. J. (1979). Motivational processes. In A. Dickinson & R. A.
Boakes (Eds.), Mechanisms of learning and motivation (pp. 171–202).
Erlbaum.
Nairne, J. S., & Rescorla, R. A. (1981). Second-order conditioning with
diffuse auditory reinforcers in the pigeon. Learning and Motivation,
12(1), 65–91. https://doi.org/10.1016/0023-9690(81)90025-4
Nasser, H. M., & Delamater, A. R. (2016). The determining conditions for
Pavlovian learning: Psychological and neurobiological considerations.
In R. A. Murphy & R. C. Honey (Eds.), The Wiley handbook on the cog-
nitive neuroscience of learning (pp. 7–46). Wiley, Ltd. https://doi.org/10
.1002/9781118650813.ch2
Norton, S. E., & Harris, J. A. (2022). Delaying extinction weakens the partial
reinforcement effect. Journal of Experimental Psychology: Animal Learn-
ing & Cognition,48(4), 321–335. https://doi.org/10.1037/xan0000319
O’Donoghue, E. M., Castro, L., & Wasserman, E. A. (2022). Hierarchical
and configural control in conditional discrimination learning. Journal of
Experimental Psychology: Animal Learning & Cognition,48(4), 370–382.
https://doi.org/10.1037/xan0000342
Pavlov, I. P. (1927). Conditioned reflexes, an investigation of the physio-
logical activity of the cerebral cortex. Oxford University Press.
Pearce, J. M. (1994). Similarity and discrimination: A selective review and
a connectionist model. Psychological Review,101(4), 587–607. https://
doi.org/10.1037/0033-295X.101.4.587
Pearce, J. M. (2002). Evaluation and development of a connectionist theory
of configural learning. Animal Learning & Behavior,30(2), 73–95.
https://doi.org/10.3758/BF03192911
Pearce, J. M., & Hall, G. (1980). A model for Pavlovian learning: Varia-
tions in the effectiveness of conditioned but not of unconditioned stim-
uli. Psychological Review,87(6), 532–552. https://doi.org/10.1037/
0033-295X.87.6.532
Perez, O. D., & Dickinson, A. (2020). A theory of actions and habits: The
interaction of rate correlation and contiguity systems in free-operant
behavior. Psychological Review,127(6), 945–971. https://doi.org/10
.1037/rev0000201
Rasmussen, A., Zucca, R., Johansson, F., Jirenhed, D. A., & Hesslow, G.
(2015). Purkinje cell activity during classical conditioning with different
conditional stimuli explains central tenet of Rescorla–Wagner model
[corrected]. Proceedings of the National Academy of Sciences of the
United States of America,112(45), 14060–14065. https://doi.org/10
.1073/pnas.1516986112
Rescorla, R. A. (1967a). Inhibition of delay in Pavlovian fear conditioning.
Journal of Comparative and Physiological Psychology,64(1), 114–120.
https://doi.org/10.1037/h0024810
Rescorla, R. A. (1967b). Pavlovian conditioning and its proper control pro-
cedures. Psychological Review,74(1), 71–80. https://doi.org/10.1037/
h0024109
Rescorla, R. A. (1968). Probability of shock in the presence and absence of
CS in fear conditioning. Journal of Comparative and Physiological Psy-
chology,66(1), 1–5. https://doi.org/10.1037/h0025984
Rescorla, R. A. (1969a). Conditioned inhibition of fear resulting from neg-
ative CS-U.S. contingencies. Journal of Comparative and Physiological
Psychology,67(4), 504–509. https://doi.org/10.1037/h0027313
Rescorla, R. A. (1969b). Pavlovian conditioned inhibition. Psychological
Bulletin,72(2), 77–94. https://doi.org/10.1037/h0027760
Rescorla, R. A. (1971). Variation in the effectiveness of reinforcement and
nonreinforcement following prior inhibitory conditioning. Learning and
Motivation,2(2), 113–123. https://doi.org/10.1016/0023-9690(71)90002-6
Rescorla, R. A. (1973a). Effect of U.S. habituation following conditioning.
Journal of Comparative and Physiological Psychology,82(1), 137–143.
https://doi.org/10.1037/h0033815
Rescorla, R. A. (1973b). Second-order conditioning: implications for theo-
ries of learning. In F. J. McGuigan & D. Lumsden (Eds.), Contemporary
approaches to learning and conditioning (pp. 127–150). Winston.
Rescorla, R. A. (1974). Effect of inflation of the unconditioned stimulus
value following conditioning. Journal of Comparative and Physiologi-
cal Psychology,86(1), 101–106. https://doi.org/10.1037/h0035964
Rescorla, R. A. (1977). Pavlovian second-order conditioning: Some impli-
cations for instrumental behavior. In H. Davis & H. M. B. Hurwitz
(Eds.), Operant Pavlovian interactions (pp. 133–164). Erlbaum.
Rescorla, R. A. (1980a). Pavlovian second-order conditioning: Studies in
associative learning. Erlbaum.
Rescorla, R. A. (1980b). Simultaneous and successive associations in sen-
sory preconditioning. Journal of Experimental Psychology: Animal
Behavior Processes,6(3), 207–216. https://doi.org/10.1037/0097-7403.6
.3.207
Rescorla, R. A. (1981a). Simultaneous associations. In P. Harzem & M.
Zeiler (Eds.), Advances in analysis of behavior (Vol. 2, pp. 47–80).
Wiley.
Rescorla, R. A. (1981b). Within-signal learning in autoshaping. Animal
Learning & Behavior,9(2), 245–252. https://doi.org/10.3758/BF03197827
Rescorla, R. A. (1982). Effect of a stimulus intervening between CS and
U.S. in autoshaping. Journal of Experimental Psychology: Animal
Behavior Processes,8(2), 131–141. https://doi.org/10.1037/0097-7403.8
.2.131
Rescorla, R. A. (1983). Effect of separate presentation of the elements on
within-compound learning in autoshaping. Animal Learning & Behav-
ior,11(4), 439–446. https://doi.org/10.3758/BF03199800
Rescorla, R. A. (1986a). Extinction of facilitation. Journal of Experimental
Psychology: Animal Behavior Processes,12(1), 16–24. https://doi.org/
10.1037/0097-7403.12.1.16
A TRIBUTE TO ROBERT A. RESCORLA 261
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
Rescorla, R. A. (1986b). Facilitation and excitation. Journal of Experimen-
tal Psychology: Animal Behavior Processes,12(4), 325–332. https://doi
.org/10.1037/0097-7403.12.4.325
Rescorla, R. A. (1987a). Facilitation and inhibition. Journal of Experimen-
tal Psychology: Animal Behavior Processes,13(3), 250–259. https://doi
.org/10.1037/0097-7403.13.3.250
Rescorla, R. A. (1987b). A Pavlovian analysis of goal-directed behavior.
American Psychologist,42(2), 119–129. https://doi.org/10.1037/0003-
066X.42.2.119
Rescorla, R. A. (1988a). Behavioral studies of Pavlovian conditioning. An-
nual Review of Neuroscience,11(1), 329–352. https://doi.org/10.1146/
annurev.ne.11.030188.001553
Rescorla, R. A. (1988b). Facilitation based on inhibition. Animal Learning
& Behavior,16(2), 169–176. https://doi.org/10.3758/BF03209061
Rescorla, R. A. (1988c). Pavlovian conditioning. It’s not what you think it
is. American Psychologist,43(3), 151–160. https://doi.org/10.1037/
0003-066X.43.3.151
Rescorla, R. A. (1989). Redundant treatments of neutral and excitatory
stimuli in autoshaping. Journal of Experimental Psychology: Animal
Behavior Processes,15(3), 212–223. https://doi.org/10.1037/0097-7403
.15.3.212
Rescorla, R. A. (1990a). Evidence for an association between the discrimi-
native stimulus and the response-outcome association in instrumental
learning. Journal of Experimental Psychology: Animal Behavior Proc-
esses,16(4), 326–334. https://doi.org/10.1037/0097-7403.16.4.326
Rescorla, R. A. (1990b). Instrumental responses become associated with
reinforcers differing in one feature. Animal Learning & Behavior,18(2),
206–211. https://doi.org/10.3758/BF03205260
Rescorla, R. A. (1990c). The role of information about the response-out-
come relation in instrumental discrimination learning. Journal of Experi-
mental Psychology: Animal Behavior Processes,16(3), 262–270.
https://doi.org/10.1037/0097-7403.16.3.262
Rescorla, R. A. (1991a). Associations of multiple outcomes with an instru-
mental response. Journal of Experimental Psychology: Animal Behavior
Processes,17(4), 465–474. https://doi.org/10.1037/0097-7403.17.4.465
Rescorla, R. A. (1991b). Associative relations in Instrumental learning:
The eighteenth Bartlett Memorial Lecture. Quarterly Journal of Experi-
mental Psychology,43B(1), 1–23.
Rescorla, R. A. (1992a). Association between an instrumental discrimina-
tive stimulus and multiple outcomes. Journal of Experimental Psychol-
ogy: Animal Behavior Processes,18(1), 95–104. https://doi.org/10
.1037/0097-7403.18.1.95
Rescorla, R. A. (1992b). Depression of an instrumental response by a sin-
gle devaluation of its outcome. The Quarterly Journal of Experimental
Psychology. B, Comparative and Physiological Psychology,44(2), 123–
136.
Rescorla, R. A. (1992c). Response-outcome vs. outcome-response associa-
tions in instrumental learning. Animal Learning & Behavior,20(3),
223–239. https://doi.org/10.3758/BF03213376
Rescorla, R. A. (1993). Inhibitory associations between S and R in extinc-
tion. Animal Learning & Behavior,21(4), 327–336. https://doi.org/10
.3758/BF03197998
Rescorla, R. A. (1994). Transfer of instrumental control mediated by a
devalued outcome. Animal Learning & Behavior,22(1), 27–33. https://
doi.org/10.3758/BF03199953
Rescorla, R. A. (1995). Full preservation of a response-outcome associa-
tion through training with a second outcome. Quarterly Journal of Ex-
perimental Psychology,48B(3), 252–261.
Rescorla, R. A. (1996a). Preservation of Pavlovian associations through
extinction. The Quarterly Journal of Experimental Psychology: Section
B,49(3), 245–258.
Rescorla, R. A. (1996b). Spontaneous recovery after training with multiple
outcomes. Animal Learning & Behavior,24(1), 11–18. https://doi.org/
10.3758/BF03198949
Rescorla, R. A. (1997a). Response-inhibition in extinction. Quarterly Jour-
nal of Experimental Psychology,50B(3), 238–252.
Rescorla, R. A. (1997b). Spontaneous recovery after Pavlovian condition-
ing with multiple outcomes. Animal Learning & Behavior,25(1), 99–
107. https://doi.org/10.3758/BF03199028
Rescorla, R. A. (1997c). Spontaneous recovery of instrumental discrimina-
tive responding. Animal Learning & Behavior,25(4), 485–497. https://
doi.org/10.3758/BF03209854
Rescorla, R. A. (2000a). Associative changes in excitors and inhibitors dif-
fer when they are conditioned in compound. Journal of Experimental
Psychology: Animal Behavior Processes,26(4), 428–438. https://doi
.org/10.1037/0097-7403.26.4.428
Rescorla, R. A. (2000b). Associative changes with a random CS-U.S. rela-
tionship. The Quarterly Journal of Experimental Psychology. B, Com-
parative and Physiological Psychology,53(4), 325–340. https://doi.org/
10.1080/713932736
Rescorla, R. A. (2001a). Are associative changes in acquisition and extinction
negatively accelerated? Journal of Experimental Psychology: Animal
Behavior Processes,27(4), 307–315. https://doi.org/10.1037/0097-7403.27
.4.307
Rescorla, R. A. (2001b). Experimental extinction. In R. R. Mowrer & S.
Klein (Eds.), Handbook of contemporary learning theories (pp. 119–
154). Erlbaum.
Rescorla, R. A. (2001c). Unequal associative changes when excitors and
neutral stimuli are conditioned in compound. The Quarterly Journal of
Experimental Psychology. B, Comparative and Physiological Psychol-
ogy,54(1), 53–68. https://doi.org/10.1080/02724990042000038
Rescorla, R. A. (2002a). Effect of following an excitatory-inhibitory com-
pound with an intermediate reinforcer. Journal of Experimental Psychol-
ogy: Animal Behavior Processes,28(2), 163–174. https://doi.org/10
.1037/0097-7403.28.2.163
Rescorla, R. A. (2002b). Savings tests: Separating differences in rate of
learning from differences in initial levels. Journal of Experimental Psy-
chology: Animal Behavior Processes,28(4), 369–377. https://doi.org/10
.1037/0097-7403.28.4.369
Rescorla, R. A. (2003a). Contemporary study of Pavlovian conditioning.
The Spanish Journal of Psychology,6(2), 185–195. https://doi.org/10
.1017/S1138741600005333
Rescorla, R. A. (2003b). Protection from extinction. Learning & Behavior,
31(2), 124–132. https://doi.org/10.3758/BF03195975
Rescorla, R. A. (2004a). Spontaneous recovery. Learning & Memory,
11(5), 501–509. https://doi.org/10.1101/lm.77504
Rescorla, R. A. (2004b). Spontaneous recovery varies inversely with the
training-extinction interval. Learning & Behavior,32(4), 401–408.
https://doi.org/10.3758/BF03196037
Rescorla, R. A. (2005). Spontaneous recovery of excitation but not inhibi-
tion. Journal of Experimental Psychology: Animal Behavior Processes,
31(3), 277–288. https://doi.org/10.1037/0097-7403.31.3.277
Rescorla, R. A. (2006a). Deepened extinction from compound stimulus
presentation. Journal of Experimental Psychology: Animal Behavior
Processes,32(2), 135–144. https://doi.org/10.1037/0097-7403.32.2.135
Rescorla, R. A. (2006b). Spontaneous recovery from overexpectation. Learn-
ing & Behavior,34(1), 13–20. https://doi.org/10.3758/BF03192867
Rescorla, R. A. (2006c). Stimulus generalization of excitation and inhibition.
Quarterly Journal of Experimental Psychology: Human Experimental Psy-
chology,59(1), 53–67. https://doi.org/10.1080/17470210500162094
Rescorla, R. A. (2007). Renewal after overexpectation. Learning & Behav-
ior,35(1), 19–26. https://doi.org/10.3758/BF03196070
Rescorla, R. A. (2008). Conditioning of stimuli with nonzero initial value.
Journal of Experimental Psychology: Animal Behavior Processes,
34(3), 315–323. https://doi.org/10.1037/0097-7403.34.3.315
Rescorla, R. A., & Colwill, R. M. (1983). Within-compound associations
in unblocking. Journal of Experimental Psychology: Animal Behavior
Processes,9(4), 390–400. https://doi.org/10.1037/0097-7403.9.4.390
262 COLWILL, DELAMATER, AND LATTAL
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
Rescorla, R. A., & Colwill, R. M. (1989). Associations with anticipated
and obtained outcomes in instrumental learning. Animal Learning &
Behavior,17(3), 291–303. https://doi.org/10.3758/BF03209802
Rescorla, R. A., & Cunningham, C. L. (1977). The erasure of reinstated
fear. Animal Learning & Behavior,5(4), 386–394. https://doi.org/10
.3758/BF03209584
Rescorla, R. A., & Cunningham, C. L. (1978a). Recovery of the U.S. rep-
resentation over time during extinction. Learning and Motivation,9(4),
373–391. https://doi.org/10.1016/0023-9690(78)90001-2
Rescorla, R. A., & Cunningham, C. L. (1978b). Within-compound flavor
associations. Journal of Experimental Psychology: Animal Behavior
Processes,4(3), 267–275. https://doi.org/10.1037/0097-7403.4.3.267
Rescorla, R. A., & Cunningham, C. L. (1979). Spatial contiguity facilitates
Pavlovian second-order conditioning. Journal of Experimental Psychol-
ogy: Animal Behavior Processes,5(2), 152–161. https://doi.org/10
.1037/0097-7403.5.2.152
Rescorla, R. A., & Durlach, P. J. (1981). Within-event learning in Pavlov-
ian conditioning. In R. R. Miller & N. E. Spear (Eds.), Information proc-
essing in animals: Memory mechanisms (pp. 81–111). Erlbaum.
Rescorla, R. A., & Freberg, L. (1978). The extinction of within-compound
flavor associations. Learning and Motivation,9(4), 411–427. https://doi
.org/10.1016/0023-9690(78)90003-6
Rescorla, R. A., & Furrow, D. R. (1977). Stimulus similarity as a determi-
nant of Pavlovian conditioning. Journal of Experimental Psychology:
Animal Behavior Processes,3(3), 203–215. https://doi.org/10.1037/
0097-7403.3.3.203
Rescorla, R. A., & Gillan, D. J. (1980). An analysis of the facilitative effect
of similarity on second-order conditioning. Journal of Experimental
Psychology: Animal Behavior Processes,6(4), 339–351. https://doi.org/
10.1037/0097-7403.6.4.339
Rescorla, R. A., Grau, J. W., & Durlach, P. J. (1985). Analysis of the
unique cue in configural discriminations. Journal of Experimental Psy-
chology: Animal Behavior Processes,11(3), 356–366. https://doi.org/10
.1037/0097-7403.11.3.356
Rescorla, R. A., & Heth, C. D. (1975). Reinstatement of fear to an extin-
guished conditioned stimulus. Journal of Experimental Psychology: Ani-
mal Behavior Processes,1(1), 88–96. https://doi.org/10.1037/0097-7403
.1.1.88
Rescorla, R. A., & Holland, P. C. (1976). Some behavioral approaches to
the study of learning. In M. R. Rosenzweig & E. L. Bennet (Eds.), Neu-
ral mechanisms of learning and memory (pp. 165–192). MIT Press.
Rescorla, R. A., & Holland, P. C. (1982). Behavioral studies of associative
learning in animals. Annual Review of Psychology,33, 265–308. https://
doi.org/10.1146/annurev.ps.33.020182.001405
Rescorla, R. A., & Skucy, J. C. (1969). Effect of response-independent
reinforcers during extinction. Journal of Comparative and Physiological
Psychology,67(3), 381–389. https://doi.org/10.1037/h0026793
Rescorla, R. A., & Solomon, R. L. (1967). Two-process learning theory:
Relationships between Pavlovian conditioning and instrumental learning.
Psychological Review,74(3), 151–182. https://doi.org/10.1037/h0024475
Rescorla, R. A., & Wagner, A. R. (1972). A theory of Pavlovian condition-
ing: variations in the effectiveness of reinforcement and nonreinforce-
ment. In A. Black & W. F. Prokasy (Eds.), Classical conditioning II (pp.
64–99). Appleton-Century-Crofts.
Rizley, R. C., & Rescorla, R. A. (1972). Associations in second-order condi-
tioning and sensory preconditioning. Journal of Comparative and Physio-
logical Psychology,81(1), 1–11. https://doi.org/10.1037/h0033333
Robbins, S. J. (1990). Mechanisms underlying spontaneous recovery in
autoshaping. Journal of Experimental Psychology: Animal Behavior
Processes,16(3), 235–249. https://doi.org/10.1037/0097-7403.16.3.235
Robbins, S. J., & Rescorla, R. A. (1989). The role of signaled periods of
nonreinforcement in responding on a random schedule in autoshaping.
Animal Learning & Behavior,17(3), 304–310. https://doi.org/10.3758/
BF03209803
Rozeboom, W. W. (1958). What is learned? An empirical enigma. Psycho-
logical Review,65(1), 22–33. https://doi.org/10.1037/h0045256
Rozin, P., & Kalat, J. W. (1971). Specific hungers and poison avoidance as
adaptive specializations of learning. Psychological Review,78(6), 459–
486. https://doi.org/10.1037/h0031878
Rusiniak, K. W., Hankins, W. G., Garcia, J., & Brett, L. P. (1979). Flavor-ill-
ness aversions: Potentiation of odor by taste in rats. Behavioral and Neural
Biology,25(1), 1–17. https://doi.org/10.1016/S0163-1047(79)90688-5
Schultz, W. (1998). Predictive reward signal of dopamine neurons. Journal
of Neurophysiology,80(1), 1–27. https://doi.org/10.1152/jn.1998.80.1.1
Seligman, M. E. (1970). On the generality of the laws of learning. Psycho-
logical Review,77(5), 406–418. https://doi.org/10.1037/h0029790
Siegel, S., & Allan, L. G. (1996). The widespread influence of the
Rescorla-Wagner model. Psychonomic Bulletin & Review,3(3), 314–
321. https://doi.org/10.3758/BF03210755
Skinner, B. F. (1938). The behavior of organisms: An experimental analy-
sis. Appleton-Century.
Speers,M.A.,Gillan,D.J.,&Rescorla,R.A.(1980).Within-com-
pound associations in a variety of compound conditioning proce-
dures. Learning and Motivation,11(2), 135–149. https://doi.org/10
.1016/0023-9690(80)90009-0
Spicer, S. G., Mitchell, C. J., Wills, A. J., Blake, K. L., & Jones, P. M.
(2022). Theory protection: Do humans protect existing associative links?
Journal of Experimental Psychology: Animal Learning and Cognition,
48(1), 1–16. https://doi.org/10.1037/xan0000314
Stafford, J. M., & Lattal, K. M. (2009). Direct comparisons of the size and
persistence of anisomycin-induced consolidation and reconsolidation
deficits. Learning & Memory,16(8), 494–503. https://doi.org/10.1101/
lm.1452209
Steinberg, E. E., Keiflin, R., Boivin, J. R., Witten, I. B., Deisseroth, K., &
Janak, P. H. (2013). A causal link between prediction errors, dopamine
neurons and learning. Nature Neuroscience,16(7), 966–973. https://doi
.org/10.1038/nn.3413
Steinfeld, M. R., & Bouton, M. E. (2022). Inhibition in discriminated oper-
ant learning: Tests of response-specificity after feature-negative and
extinction learning. Journal of Experimental Psychology: Animal Learn-
ing & Cognition,48(4), 413–434. https://doi.org/10.1037/xan0000337
Stout, S. C., & Miller, R. R. (2007). Sometimes-competing retrieval
(SOCR): A formalization of the comparator hypothesis. Psychological
Review,114(3), 759–783. https://doi.org/10.1037/0033-295X.114.3.759
Swartzentruber, D., & Rescorla, R. A. (1994). Modulation of trained and
extinguished stimuli by facilitators and inhibitors. Animal Learning &
Behavior,22(3), 309–316. https://doi.org/10.3758/BF03209839
Testa, T. J. (1975). Effects of similarity of location and temporal intensity
pattern of conditioned and unconditioned stimuli on the acquisition of
conditioned suppression in rats. Journal of Experimental Psychology:
Animal Behavior Processes,1(2), 114–121. https://doi.org/10.1037/
0097-7403.1.2.114
Tobler, P. N., Dickinson, A., & Schultz, W. (2003). Coding of predicted
reward omission by dopamine neurons in a conditioned inhibition para-
digm. The Journal of Neuroscience,23(32), 10402–10410. https://doi
.org/10.1523/JNEUROSCI.23-32-10402.2003
Tolman, E. C. (1932). Purposive behavior in animals and men. Century/
Random House U.K.
Trapold, M. A., & Overmier, J. B. (1972). The second learning process in
instrumental learning. In A. H. Black & W. F. Prokasy (Eds.), Classical
conditioning II: Current theory and research (pp. 427–452). Appleton-
Century-Crofts.
Uengoer, M., Lachnit, H., & Pearce, J. M. (2020). The role of common ele-
ments in the redundancy effect. Journal of Experimental Psychology:
Animal Learning and Cognition,46(3), 286–296. https://doi.org/10
.1037/xan0000236
A TRIBUTE TO ROBERT A. RESCORLA 263
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
Waelti, P., Dickinson, A., & Schultz, W. (2001). Dopamine responses
comply with basic assumptions of formal learning theory. Nature,
412(6842), 43–48. https://doi.org/10.1038/35083500
Wagner, A. R. (1981). SOP: A model of automatic memory processing in
animal behavior. In N. E. Spear & R. R. Miller (Eds.), Information proc-
essing in animals: Memory mechanisms (pp. 5–44). Erlbaum.
Wagner, A. R. (2008). Evolution of an elemental theory of Pavlovian con-
ditioning. Learning & Behavior,36(3), 253–265. https://doi.org/10
.3758/LB.36.3.253
Wagner, A. R., & Brandon, S. E. (1989). Evolution of a structured connection-
ist model of Pavlovian conditioning (AESOP). In S. B. Klein & R. R.
Mowrer (Eds.), Contemporary learning theories: Pavlovian conditioning
and the status of traditional learning theory (pp. 149–189). Erlbaum.
Wagner, A. R., & Brandon, S. E. (2001). A componential theory of Pav-
lovian conditioning. In R. R. Mowrer & S. B. Klein (Eds.), Contempo-
rary learning: Theory and applications (pp. 23–64). Erlbaum.
Wagner, A. R., Logan, F. A., Haberlandt, K., & Price, T. (1968). Stimulus
selection in animal discrimination learning. Journal of Experimental
Psychology,76(2), 171–180. https://doi.org/10.1037/h0025414
Wagner, A. R., & Rescorla, R. A. (1972). Inhibition in Pavlovian condi-
tioning: application of a theory. In R. A. Boakes & S. Halliday (Eds.),
Inhibition and learning (pp. 301–336). Academic Press.
Wasserman, E. A., & Castro, L. (2022). The Rescorla-Wagner model: The
culmination of Hume’s theory of causation. Journal of Experimental
Psychology: Animal Learning & Cognition,48(4), 315–320. https://doi
.org/10.1037/xan0000325
Zimmer-Hart, C. L., & Rescorla, R. A. (1974). Extinction of Pavlovian
conditioned inhibition. Journal of Comparative and Physiological Psy-
chology,86(5), 837–845. https://doi.org/10.1037/h0036412
Received September 16, 2022
Accepted September 16, 2022 n
Call for Nominations
The Publications and Communications (P&C) Board of the American Psychological Association
has opened nominations for the editorships of Clinician’s Research Digest;Psychology, Public
Policy, and Law;Psychology and Aging;Professional Psychology: Research and Practice;Journal
of Experimental Psychology: Learning, Memory, and Cognition; and the Journal of Personality and
Social Psychology: Interpersonal Relations and Group Processes. Marisol Perez, PhD, Michael E.
Lamb, PhD, Elizabeth A. L. Stine-Morrow, PhD, Kathi A. Borden, PhD, Aaron S. Benjamin, PhD,
and Colin Wayne Leach, PhD, respectively, are the incumbent editors.
Candidates should be members of APA and should be available to start receiving manuscripts in
early 2024 to prepare for issues published in 2025. The APA Journals program values equity,
diversity, and inclusion and encourages the application of members of all groups, including women,
people of color, LGBTQ psychologists, and those with disabilities, as well as candidates across all
stages of their careers. Self-nominations are also encouraged.
Search chairs have been appointed as follows:
Clinician’s Research Digest, Chair: Michael Roberts, PhD
Psychology, Public Policy, and Law, Co-Chairs: Pamela Reid, PhD, and Hortensia Amaro, PhD
Psychology and Aging, Chair: Mo Wang, PhD
Professional Psychology: Research and Practice, Chair: Mark Sobell, PhD
Journal of Experimental Psychology: Learning, Memory, and Cognition, Chair:
Steve Kozlowski, PhD
Journal of Personality and Social Psychology: Interpersonal Relations and Group Processes,
Chair: Richard Petty, PhD
Nominate candidates through APA’s Editor Search website (https://editorsearch.apa.org).
Prepared statements of one page or less in support of a nominee can also be submitted by email to
Jen Chase, Journal Services Associate (jchase@apa.org).
Deadline for accepting nominations is Monday, January 9, 2023, after which phase one vetting will
begin.
264 COLWILL, DELAMATER, AND LATTAL
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Content may be shared at no cost, but any requests to reuse this content in part or whole must go through the American Psychological Association.
