Printer Friendly


The Decline of Classical Behaviourist Learning Theory

Understanding the process through which individuals learn and how behaviour and cognition may change as a consequence has generated much research, debate and scholarship. Early studies postulated that there is an inextricable link between learning and behaviour. In this respect, it is interesting to notice that in the 1960s, learning was thought to be directly related to change in behaviour (Mackintosh, 1997; 2009). Moreover, one can trace this link between behaviour and learning back to the work of Pavlov and Skinner on classical and operant conditioning respectively. Their theoretical assumptions informed much of classical learning theory. More specifically, learning was essentially regarded as an end product of some processes. In this way, research within the experimental psychology domain was characterised by its main focus on animal models derived from conditioning (stimulus-response) research. Such models are grounded on some Darwinian principles which postulate that organisms are inherently selfish, competitive and dependent on what is necessary for the species to survive. From this vantage point, behaviourists like Skinner aligned himself with Darwin's ideas and postulated that animals are born with a mind in a 'blank' state and that their learning and survival is a consequence of simple environmental adaptation. This is precisely the starting point of Mackintosh criticism, discussion, and analysis. Mackintosh (1997) argued that mental activity and human behaviour is far too complex to be understood in terms of highly reductionist behavioural experiments.

Thus, given the behaviourist paradigm failure to account for complexities surrounding necessary and sufficient conditions for the occurrence of successful conditioning and learning, there has been a paradigm shift from behaviourism to cognitivism (McLaren and Mackintosh, 1989). It follows that Mackintosh also exposed both behaviourism and cognitivism for failing to show adequate generalizability in human behaviour. In this respect, it would appear that, despite most experiments providing evidence for both operant conditioning and Pavlovian conditioning, these experiments were simply based on animals and their behaviour. There is an evident problem in assuming that general laws which are characteristic of animal behaviour can somehow describe the much more complex pattern of human, thought, behaviour and mental processes. Mackintosh also subtly touches upon the issue of completely dismissing mental processes as solely a product of stimulus-response dynamics and the lack of engagement in understanding cognitive processes shown by radical behaviourists.

Thus, Mackintosh (1997) article begins with a highly selective analysis and discussion of the most prominent historical origins of this research area and the way in which learning-performance distinction may in fact provide a useful framework for current and future theoretical analysis. Mackintosh goes on to discuss the extent to what instrumental and classical conditioning are actually separate processes. He, therefore, asserted that a common feature between them is the concern of defining the 'unit of learning' and the nature of the associative structure that results from animals being exposed to certain conditioning procedures. In this way, he essentially emphasises the importance of establishing a basic independent variable whilst at the same time he shows an unquestioning acceptance of associative learning principles. Thus, he essentially explores the nature of the fundamental independent variable for conditioning procedures. More specifically, he placed particular emphasis on conditioning procedures which are particularly crucial for studying choice. In this respect, it is also interesting to notice that Mackintosh situates such procedures under a more general umbrella of discrimination learning. Moreover, he asserted that behaviour is distributed through some sort of governing rule even after establishing the associative strengths of individual alternatives.

Mackintosh considers two interesting possibilities; the "matching" rule (whereby the animal behaviour is generated probabilistically in proportion to its relative strength), and the maximisation rule, in which animals make a forced choice based upon which alternative appears to have greater strength. One should also notice that Mackintosh deliberately refuses to give any careful scrutiny to previous quantitative evidence relating to the aforementioned theoretical stance. In this way, he simply dismisses such evidence as being too vague, and simply being confounded by a vast array of different topics which do not seem to have any relation to each other. Nonetheless, he shows some signs of academic maturity and depth of analysis by admitting that problems with existing quantitative evidence are rather complicated ones, and that organising and reorganising contradictory claims within relevant literature is by no means a trivial task.

It also appears that Mackintosh shows little or no appreciation for previous claims postulating that "all behaviour is a choice". In this respect, he argued that when deciding whether to respond or not, animals' choice cannot always be described by some rule of choice. In addition, one of the strengths of this analysis is the way in which he regarded such conditioned responses as not being bound to, or indeed being a manifestation of quantitative properties of presumed 'governing rules' of animal behaviour. It follows that, one of the strengths of this analysis is the fact that it places much emphasis on the idea that occurrence of behaviour is regarded as a product of previous reinforcement history, as well as a consequence of operations affecting the value of the reinforcer. Thus, he also assumes that the concept "response strength" is simply a result of the value of the reinforcer in conjunction with strength of the association between the response and the reinforcer. It is precisely in this manner that his theoretical construct "association" emerges and it systematically reorganises a large body of knowledge into a coherent whole.

It should be apparent that one of the weaknesses of Mackintosh thesis is that his division between associative learning stance and behaviour analysis may not be underpinned simply by differences in their subject matters. Despite the fact that behaviour analysis and associative learning tradition substantially diverge in their accounts, it is still a major challenge to create some interaction between the two theoretical stances in animal learning. In summary, one should notice that one of the most pervasive similarities between associative learning tradition and behavioural analysis theoretical approaches has been the fact that both have attempted to comprehend the nature of associative structures mediating conditioning effects. In turn, Mackintosh synthesised associative learning theory in terms of its implicit argument which favours associative concepts. More specifically, he regards such associative concepts as prerequisites for bringing coherence to current knowledge relating to the effects of conditioning procedures. Furthermore, perhaps one of the major criticisms of Mackintosh's paper is his complete disregard for any theoretical perspective which has evolved from the Skinnerian research tradition. Yet, paradoxically, one of the strengths of Mackintosh thesis is the fact that he does this with much theoretical sophistication.

It appears that Mackintosh is perhaps too dismissive of the behaviourist account and its previous findings. However, despite much scepticism, one should take into account that classical behaviourist epistemology provided useful insights into understanding the role played by environmental settings upon human behaviour. More specifically, it provided useful insights into understanding the nature of substance addiction and related psychotherapeutic interventions for substance abuse (Alcaro, Huber & Panksepp, 2007). In addition, it could also be argued that behaviourist principles are deeply embedded within the learning and education and societal domains. This may particularly be the case when one is dealing with reward and punishment, whereby certain types of behaviour are encouraged and other (which are deemed counterproductive) are discouraged (Alcaro, Huber & Panksepp, 2007). It should also be evident that radical behaviourist assumptions may have attracted much hostility from a considerable number of scholars. Perhaps, behaviourism itself was not necessarily the problem, but instead the radical approach adopted by theorists and researchers. In this way, one should note that there is always a danger of slipping into some sort of "scientific fundamentalism". Within the academic arena there is often the prevalence of culture of certainty and 'blind' theoretical commitment. Thus, whether behaviourism or cognitivism, when conducting research one should always take into account that knowledge is never complete or finished. Instead, knowledge evolves in proportion to one's understanding of the world, given that no one version of the truth is completely absolute or infallible. In turn, the subject matter of psychology evolves and changes as a function of the complexities and challenges presented during our endeavour to understand the nature of mind, thought and behaviour (De Grandpre, 2000). Therefore, behaviourist approach to learning should not be completely dismissed given that it acted as a catalyst for future developments within experimental Psychology.

The Importance of Associative Learning

Over that past two decades, the associative learning tradition perhaps enjoyed a very productive period, whilst the behaviour analysis theoretical stance has witnessed somewhat of a stagnant period. Whilst advocating that the associative learning theory is well alive, Mackintosh (1997) argued that theorists, who subscribe to this theoretical stance, take a much nuanced and sophisticated stance when it comes to understanding the necessary and sufficient conditions for the occurrence of successful conditioning. Associative learning takes into account a wide range of factors involved in the process of learning including abstract concept learning, transitive inference and analogical reasoning. According to Mackintosh these factors must be taken into account when developing a suitable theory of the mind.


In conclusion, it should be evident that Mackintosh article is a highly valuable contribution to the body of knowledge in animal learning. His authoritative assessment of associative learning theory provides useful insights into how one might synthesize the major conclusions connected to recent findings within this research area. Perhaps one of the major criticisms of Mackintosh's paper is his complete disregard for any theoretical perspective which has evolved from the Skinnerian research tradition. In his paper, Mackintosh (1997) addresses an important question about the extent to what changes in behaviour of animal experiments could in fact reflect changes in perception.

Despite some limitations of classical behaviourism as a comprehensive theory, one could argue that it is still valid in certain areas of Psychology. For instance, as discussed previously certain ideas from classical behaviourist theories have been applied to the treatment of neurosis and phobias (Schwartz & Levy, 1982). Similarly, Pavlovian principles have also been applied to the treatment of addiction, given that addiction occurs because of both the pleasurable physiological effects of addictive substances such as nicotine and alcohol. In this way, classical behaviourism should not be completely dismissed given that it might have been a small piece in much wider 'puzzle' of theory of mind. Furthermore, classical behaviourism, and behaviour analysis provided the prerequisites for Mackintosh criticism, analysis, and discussion of the importance of developing and adopting a much nuanced and sophisticated theory (such as associative learning).


Alcaro, A., Huber, R., & Panksepp, J. (2007). Behavioral functions of the mesolimbic dopaminergic system: an affective neuroethological perspective. Brain Research Reviews, Vol.56, (2), pp.283-321.

De Grandpre, R. J. (2000). A science of meaning: Can behaviorism bring meaning to psychological science? American Psychologist, Vol.55, (7), pp.721-750

Mackintosh, N. J. (1997). Has the wheel turned full circle? Fifty years of learning theory, 1946-1996. The Quarterly Journal of Experimental Psychology: Section A, 50(4), 879-898.

Mackintosh, N.J. (2009) Varieties of perceptual learning. Learning and Behavior, Vol.37, pp.119-25

McLaren, I.P.L., Kaye, H. and Mackintosh, N.J. (1989). An associative theory of the representation of stimuli: applications to perceptual learning and latent inhibition. In R.G.M. Morris (Ed.) Parallel Distributed Processing -Implications for Psychology and Neurobiology. Oxford. OUP

Schwartz, B. S., Levy, G. A., Fair, D. S., & Edgington, T. S. (1982). Murine lymphoid procoagulant activity induced by bacterial lipopolysaccharide and immune complexes is a monocyte prothrombinase. The Journal of Experimental Medicine, Vol.155, (5), pp.1464-1479.

Mauro Ramos De Jesus Pereira

Lund University Department of Psychology
COPYRIGHT 2016 Oxford Mosaic Publications Limited
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2016 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:De Jesus Pereira, Mauro Ramos
Publication:Journal of Social and Psychological Sciences
Date:Jul 1, 2016

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters