In this blog post I will discuss the fears of humans being controlled by behavioural science. This fear receded as criticisms of behaviourism made the discipline seem less capable of achieving the bombastic claims made by the likes of Skinner and Watson. But behaviourism has advanced well beyond Skinner and Watson and current behavioural science is again looking at ways of controlling rule governed behaviour. After discussing Skinner’s early work on control and counter control I will evaluate contemporary behavioural science and its potential for predicting and controlling verbal behaviour linking behavioural control with the rise of Artificial Intelligence.

In Skinner’s ‘Science and Human Behaviour’ in 1953 he speculated on ways behavioural technology could improve the lives of people if implemented properly. However, there were those who didn’t like the imposition of scientists into human affairs. Some of this opposition was simple fear that science was encroaching on subject matters typically tackled by the humanities. But for others they found the Skinner’s emphasis on “prediction and control” chilling. People feared a future where behavioural scientists would eventually achieve control over ordinary human subjects and shape and control their behaviour at the whim of those in power.

            However, as we learned more about animal behaviour, and the factors impinging on such behaviour, the threat from behavioural science began to seem less and less. Breland and Breland’s (1963) work on animal training demonstrated that instinctual drift would mean that non-human animals’ behaviour wasn’t as malleable as earlier naïve behaviourists thought. Skinner, who had long stressed both phylogenetic and ontogenetic factors playing a role in animal behaviour welcomed the work of the Breland’s. On Skinner’s way of thinking it was the behaviourists job to discover the different ways behaviour could be shaped and controlled through different schedules of reinforcement. The behaviourist wasn’t in the game of stipulating how malleable different organisms were. Nonetheless, despite the Breland’s work being congenial to Skinner’s behaviourism, for the public, instinctive drift made the thoughts of behaviourists gaining control over people and shaping their behaviour seem less threatening.

             As behavioural scientists continued to study human’s operating under schedules of reinforcement there was more reason to think that humans couldn’t be just shaped at a whim through schedules of reinforcement. Hundreds of behavioural studies on rule governed behaviour[1], have demonstrated that when humans were operating under rules this made them less sensitive to the contingencies of reinforcement. Children below the age of 5 could be shaped under schedules of reinforcement in a similar way to a rat, but once they passed 5 and could follow rules their behaviour became less sensitive to the contingencies of reinforcement (Bentall 1985).

            Like the Breland’s work, work on rule-following sprung up in behaviourism (Skinner 1963) and demonstrated that human behaviour was more complex than the behaviour of other organisms. When Sidman (1971) experimentally demonstrated that humans could derive untrained stimulus equivalence, and Hayes (1989) demonstrated that humans could derive untrained relational frames (coordination, comparison, hierarchy, etc). The human under behavioural science began to closer resemble the human as described by cognitive scientists (built with innate constraints, have a species-specific capacity for productive reasoning etc), than it did the human as described by early behaviourists.

            With all these facts in place people in the popular press stopped worrying about nefarious behaviourists controlling them. The stuff of nightmares in the popular imagination is now neuroscientists controlling us, or AI controlling us. Fear of behaviourists controlling us is now a part of our quaint past on a par with fears in the nineteen fifties fearing visitors from the Moon.

            Nonetheless, Skinner’s point still stands. Our behaviour is controlled by facts in our ontogeny and phylogeny as well as a variety of other factors. Our behaviour is always controlled and science, which gives us the best tools to predict and control phenomenon can help us as a species, gain some degree of control of our collective behaviour. Some of this control is welcome, we are glad that people follow rules that forbid them to enter a stranger’s house without permission or forbid them to drive on the wrong side of the road.

            But when it comes to things like Verbal Behaviour people are particularly reticent in ceding control to scientists. Our ability to talk and think about whatever we want is a key component of what it means to be human. Thus, it is the stuff of our nightmares that scientists or politicians gain the ability to control what we can say. Famously, in Orwell’s dystopian novel 1984 the party controlling society, had banned the use of certain words such as ‘Freedom’ and ‘Equality’ etc. With the aim being that to the extent that thought depends on language if you can ban the words from public consciousness people will no longer be able to think the thoughts. Compelling evidence from cognitive science indicates that it is not necessary to have a word in public language in order use a concept (Pinker 1994, 2002). While cognitive science has conclusively shown that our thought is not entirely determined by our language. It is evident that our thought processes are shaped by the type of language we acquire. Lakoff (1987), has demonstrated how the metaphors we adapt will influence the way we think about the world. So, the fear still exists that nefarious scientists in conjunction with political elites will some how conspire to control what we can say perhaps even influencing what we can think. Hence, in recent years there has been an explosion of public intellectuals; usually podcasters who argue that there should be no limits on who can say what. These people call themselves free-speech absolutists.

            But as of yet the science of verbal behaviour doesn’t demonstrate any real capacity to control Verbal Behaviour. In fact, as discussed above, once a creature becomes verbal and starts following rules their behaviour becomes less sensitive to the contingencies of reinforcement. Furthermore, as linguists like to note humans appear to have the capacity to construct new sentences which have never been spoken before using their ability to use productive syntax. This type of creativity doesn’t appear to be under the control of any kind environmental or behavioural control.

            Nonetheless, we known anecdotally that the type of sentences we use are to some degree shaped because of our history of interaction with particular groups. Skinner tells an anecdotal story about meeting his parents and his lecturers at his graduation. He found the conversation difficult to navigate because he had a history of reinforcement for talking to his family in a particular manner, and a different history of reinforcement for talking in a different manner to his professors. In the conversation he felt the pull of two competing verbal repertoires as he interacted with the two groups of people.

            Obviously, however just because Skinner has verbal repertories he used when talking to people in different groups, he was nonetheless capable of creating new sentence to talk about other things which may have interested him or his interlocuters. There is no reason to think that his verbal repertoires were entirely determined by past reinforcement history. Nonetheless, the fact that he felt the pull of two competing verbal repertoires does indicate that the type of speech acts we engage in are to some degree shaped by social contingencies.

            Behaviourists have long noted that humans aren’t just passive organisms who submit to control. Humans typically don’t like being controlled and respond to control with two key tactics (1) Aggressive Responses, (2) Moving out of the range of range of controllers or passive resistance (Skinner 1953 p. 193). But in an unequal society with one group being more powerful than another group; aggressive responses may be pointless. Furthermore, with the ability to increase of technology into every aspect of our lives it is more and more difficult to move out of the range of controllers. Humans spend an inordinate amount of time online where they can be reached by a variety of different forces.

            As we discussed above when humans begin to develop their language and start following verbal rules this makes them less sensitive the contingencies of reinforcement. Hence, their behaviour when rule governed is more difficult to behaviourally control. Behaviourists have noted that humans engage in a type of rule governed behaviour which is shaped by social consequences. They call this Pliance. There is rule governed behaviour which is controlled by tracking. Where the rules are adopted because they accurately track environmental facts. And augmentative rule following which alters the extent to which rule governed behaviour is reinforcing.

Importantly for our purposes they have started to develop techniques to help them control rule governed behaviour. In a recent paper behaviourists Spencer et al (2022) noted that people had trouble following verbal rules for three reasons: (2) Lack of Credibility of the speaker, (2) Lack of the ability of the speaker or authority figure to mediate contingencies of rule-following, (3) implausibility of a given rule. And they argue that these facts can be overcome through the following techniques. Strengthening augmental control by connecting the rule to what the individual values. Monitoring rule following to assess Pliance rates. Ensuring the contingencies supporting rule following track reality accurately. Deemphasizing freedom threatening language (Spencer et al 2022).

As things stand behavioural techniques employed by psychologists in relation to things like tracking and Pliance typically relate to therapeutic settings. Thus, in therapy a therapist may discuss with a patient how their verbal repertoire is geared toward social compliance sometimes at the expense of tracking reality accurately. But as accuracy in understanding rule following behaviour and how it is controlled increased it may eventually be used by governments to shape how people verbally respond to societal rules. A step in this direction was Stapelton et al 2022 which explored the rule following behaviour of people during the Covid-19 pandemic.

With the creation of Large Language Models which will increasingly appear online it will be possible that most “people” you interact with online will be Large Language Models or other kinds of AI. Large Language Models when being trained can be shaped to reflect the values of the person designing it. Thus, during training using reinforcement the Models will be trained to only answer in ways consistent with the biases of the creators. Children growing up who may be perennially online could end up receiving a massive amount their linguistic data from Large Language Models.

These models sometimes hallucinate answers based on statistical expectations. And other times they will be trained to reflect the values of their designers. It is more and more possible to track what people’s values are based on their online behaviour. What they look at; what they buy, who they interact with. If Large Language Models are linked to attractive avatars and given their impressive manner of being able to answer a surprising number of questions. People may start viewing them as authorities. And this is the first step in them being used as reliable givers of information. Children growing up could end up following the rules of unconscious LLMs whose values were shaped by their creators. This wouldn’t be verbal control of the type Orwell worried about. But it would be a way of shaping the values and verbal repertories of people in ways we may be entirely unaware of.

[1] See for example: Weiner et al (1964), Lippman and Myer (1967), Lowe et al (1983), Hayes et al (1986).

The Setting

The act of referring to objects, so effortlessly performed in everyday discourse, belies the complex web of philosophical inquiry it has spurred since the dawn of analytic philosophy. In the annals of this intellectual tradition, a recurring motif emerges: the relentless quest to mend the perceived fractures in the relationship between words and the world they represent. As the century unfolded, a dichotomy crystallized within the realm of linguistic philosophy. On one side stood the cognitivists, disciples of Chomsky, who espoused the view of language as an internal computational apparatus for thought. According to Chomsky’s doctrine, language itself did not possess inherent reference but served as a conduit for expressing thoughts about the external realm. On the opposing front were the behaviourists, led by Skinner, who conceptualized language as verbal behavior moulded by environmental contingencies. Skinner’s contention that the traditional notion of reference was dispensable gave rise to the notion of the “tact,” a verbal operant subject to control by non-verbal stimuli and shaped by the three-term contingency.

Despite their discordant philosophies, both camps found common ground in their dismissal of the reference relation as dissected by analytic philosophers. This scepticism toward linguistic philosophy’s pursuits of clarity and repair continues. But work over the last half a century has laid the groundwork for some convergence between analytic philosophy and behavioural science. In this exploration, we delve into the fertile terrain where analytic philosophy and behavioural science intersect, focusing on the burgeoning field of Relational Frame Theory (RFT). Through an examination of the perspectives of Quine and relational frame theorists, we embark on a journey to unravel the intricacies of observation sentences—how they are acquired, developed, and intertwined with our perception of the world. Join us as we navigate the intricate labyrinth of language and cognition, seeking to illuminate the elusive nature of referentiality and its implications for our understanding of the human mind.

              The Evolution of language

In their groundbreaking work, Hayes and Sanford (2014) delve into the evolutionary origins of our verbal behavior, painting a vivid picture of how our linguistic abilities evolved over time. They dissect our communication skills, drawing parallels between humans and other non-verbal creatures, highlighting shared traits such as vocalizations to influence behavior and social referencing. Arguing from the vantage point of group selection, Hayes and Sanford suggest that our cooperative instincts played a pivotal role in shaping our linguistic capacities. They illustrate this through a compelling scenario: two individuals near an apple tree, both equipped with the word “apple.” When one person vocalizes the word and the fruit is out of reach, the other steps in to retrieve it, fostering cooperation and mutual understanding. This cooperative dynamic, they contend, marks the genesis of our ability to engage in a frame of coordination, where words seamlessly align with their referents. Through reinforcement and shared experiences, individuals develop a nuanced understanding of the relationship between sounds and objects, paving the way for combinatorial entailment and the intricate web of relational frames. In essence, Hayes and Sanford propose that our cooperative instincts underpin our capacity to frame and derive meaning from language, ushering in a new era of understanding in the evolutionary origins of human communication.

Once this frame of coordination was acquired humans would then have the capacity to recognize mutuality in a frame. And they argue that repeated application of mutuality would give an organism the ability to use combinatorial entailment (ibid p.123). Thus, on this conception the capacity to relationally frame is created primarily by our cooperative instinct.

   Individual or Shared Behavioural Streams

In the intricate narrative spun by Sanford and Hayes (2014), the spotlight falls on the profound interplay between human cognition and cooperative instincts, casting a revealing light on the evolution of verbal behavior. Central to their thesis is the notion of coordination between words and objects, a feat achieved through social derivation and contextual control over the “is” relation. Illustrating their argument with the allegory of two individuals wielding the word “apple” as a Mand, Sanford and Hayes beckon us into a world where language becomes a conduit for shared experiences and mutual understanding. Yet, lurking beneath the surface lies a critique of individualistic interpretations, epitomized by Barnes-Holmes (2000), who dissected such instances of triangulation within the confines of individual behavioural streams. In this intricate dance of cognition and cooperation, the tale takes a twist as the narrative shifts to the realm of joint attention. Here, the shared object of experience emerges as the linchpin, transcending individual streams and fostering a communal understanding. While behavioural pragmatists grapple with the ontological underpinnings, their reluctance to delve into intersubjective agreement leaves a void in their narrative, one that can only be filled by acknowledging the symbiotic relationship between individual streams and communal experiences. Thus, as the story unfolds, Sanford and Hayes beckon us to embark on a journey through the labyrinth of human cognition, where language and cooperation intertwine to shape our collective understanding of the world.

         Cometh the Hour Cometh the Man

Enter Quine, the philosophical voyager navigating the treacherous waters of linguistic ambiguity and cognitive convergence. Armed with intellect and inquiry, Quine set out to unravel the enigma of shared understanding amidst the cacophony of sensory receptors. In Quine’s labyrinth of inquiry, the conundrum of observation sentences loomed large, casting shadows of doubt upon the very fabric of linguistic communion. While his behavioural contemporaries grappled with the intricacies of Mand and Tact, Quine delved into the depths of epistemological and semantic inquiry, seeking to unveil the secrets of meaning and consensus. Yet, amidst the philosophical fog, parallels emerged between Quine’s quest and the behavioural pragmatists’ plight. Both faced the daunting task of reconciling individual streams of experience with the communal tapestry of cognition, each seeking solace in the embrace of naturalistic explanation. Quine’s solution, an appeal to perceptual harmony and empathetic resonance, stood as a beacon of enlightenment amidst the turbulent seas of inquiry. Through the lens of empathy, Quine discerned the threads of joint attention and social referencing, weaving a tapestry of understanding that transcended individual perspectives. And yet, the tension persisted between Quine’s ontological aspirations and the pragmatists’ yearning for epistemological clarity. As Barnes-Holmes et al. posited the notion of separate behavioural streams, Quine’s empathy-driven paradigm stood as a testament to the interconnectedness of human cognition. In the end, Quine’s journey served as a testament to the intricate dance of intellect and instinct, weaving a narrative of inquiry and insight that continues to resonate across the landscape of philosophical discourse.

                           Introduction

David Barnes-Holmes (2000) in his ‘Behavioural Pragmatism: No Place for Reality and Truth’ criticized Quine for offering a non-solution to the problem of homology. The problem of homology sprung up when Quine was trying to explain how different subjects could assent or dissent from observation sentences. Quine tried to cash out the stimulus meaning of observation sentences in terms of the triggering of sensory receptors. But this led to a difficulty of explaining how two people whose sensory receptors could be triggered differently could converge on stimulus meanings. Quine’s non-solution was that the scientist in practice could ignore the problem because it wouldn’t affect their experiments (Quine 1974 p. 24). Barnes-Holmes argued that while Quine was counselling to ignore the problem based on pragmatic principles contextual behaviourists could remove the problem altogether by adopting a consistent pragmatist approach (Barnes-Holmes 2000 p. 197).

                Quine’s Darwinian Solution

            But Barnes-Holmes didn’t note that Quine (1996) had already come up with a different solution to the problem of homology which didn’t involve merely ignoring it on pragmatic grounds (Barnes-Holmes 2000 p. 196). Quine’s original definition of stimulus-meaning as being provided by the triggering of sensory receptors was difficult to sustain because of his publicity criterion in relation to observation sentences (Kemp 2019 p. 60). Given that people’s sensory receptors could be triggered in a myriad of different ways there is no real justification for assuming shared sensory triggering’s are responsible for our agreement on whether to assent or to dissent from an observation sentence at all.

            Quine though thought that shared stimulus meanings could be accounted for by assuming that humans have a shared perceptual similarity space which underlay their grouping of things together (Quine 1996 p. 161). This similarity space would be partly innate though modifiable through training (Quine 1995 p. 21). He argued that this perceptual similarity space would have survival value and hence would be passed on through natural selection. Put in Skinnerian terminology Quine speculated that the phylogenetic contingencies of survival would be responsible for our perceptual similarity space. And he argued that the nature of these similarity spaces could be determined empirically:

“But perception for all its mentalistic overtones, is accessible to behavioural criteria. It shows itself in conditioning responses. Thus, suppose we provide an animal with a screen to look at and a lever to press. He finds that the pressed lever brings a pellet of food when the screen shows a circular stripe, and that it brings a shock when the screen shows four spots spaced in a semi-circular arc. Now we present him with these same four spots, arranged as before, but supplemented with three more to suggest the complementary semicircle. If the animal pressed the lever, he may be said to have perceived the circular Gestalt rather than the component spots. (Quine 1974 p. 4)

All this is well within the remit of typical animal studies.

Quine was bringing in the notion of a partially innate perceptual similarity shared amongst humans to explain how we converge on the same stimulus meanings for observation sentences. And observation sentences were obviously meant as our empirical check point for our theory of the world. The purpose of this was to give us a naturalistic theory of how we go from Stimulus to Science. And shared innate perceptual similarities cashed out in terms of natural selection and neural structures is sufficiently objective for Quine to rest satisfied with the explanation.

    Barnes-Holmes Behavioural Pragmatism

Barnes-Holmes never mentioned Quine’s Darwinian attempt of a solution to the problem of homology. So obviously he never framed any reply. Before, considering what he may think of Quine’s attempted solution I will first briefly outline his own behavioural pragmatist solution to the homology problem and then relate his attempted solution to Quine’s Darwinian Solution.

 When outlining his behavioural pragmatism Barnes-Holmes notes that it relies on three key assumptions to justify the position. Assumption 1: What is known is always a behavioural function. Assumption 2: The activity of each organism participates in a different behavioural stream. Assumption 3: The activity of the behavioural pragmatist participates in a behavioural stream (Ibid p. 198).

            Each of these assumptions are extremely controversial. Assumption 1 argues that what is known is always a behavioural function not a behaviour independent reality. It isn’t particularly controversial to state that we know things through a function of our behaviour. Barnes-Holmes gives the example of an apple and states that we know the apple through its functions such-as how it elicits stimulations like salivation, or it being a discriminative stimulus for Verbal Behaviour such as ‘this is an apple’, or as reinforcing stimulus for saying something like ‘give me an apple’ (Ibid p. 197). The few examples he gives involve the ontogenetic interaction of a human subject with an apple. He doesn’t mention any phylogenetic factors which would go into the human’s behavioural interaction with the apple. Though presumably he would admit that phylogenetic factors may play such a role. Skinner long stressed that some phylogenetic factors will play a role in an organism’s behaviour (Skinner 1974 p. 228). Likewise, Hayes and Sanford (2014) suggest that to ensure that asking for an apple will have reinforcing consequences we will have to assume a cooperative instinct.

            None of this will be particularly controversial. There are empirical details to fill in about the nature of our behavioural interactions with entities such as apples. And debates about the role of perception as opposed to brute behaviour in our knowledge of things like apples. But depending on how we construe behaviour it is eminently sensible to suggest that we only know something though our behavioural functions with it. But Barnes-Holmes goes further than this. He isn’t content to claim we only know something through behavioural interactions with it; he goes on contrast this with a belief in a physical apple at all:

“In commonsense terms, the apple is a physical thing that exists independently of behavior. For the behavioural pragmatist, however, the apple is defined only in terms of its behavioural functions that emerge in a particular stream of behavioural interactions.” (Ibid p. 197)

In the above quote the notion of a physical thing which exists independent of behaviour is rejected. He isn’t claiming that a mind independent apple doesn’t exist; rather he merely states that the behavioural pragmatist’s definition of apple doesn’t rely on notions of physical objects that exist independent of behaviours. Though he does go on to say that behavioural pragmatist will sometimes talk as if some objects exist independent of behaviour (Ibid p. 198). But such talk involves no ontological commitment as Verbal Behaviour in the technical sense doesn’t ‘refer’ or ‘correspond’ to an external reality (Ibid p. 199).

            It is important to be clear about the commitments they are making. When Barnes-Holmes is making claims about defining objects in terms of their behavioural functions the examples he gives of behaviour are of classical conditioning where an object elicits salivation in an organism, discriminative stimulus which increases the probability of a particular tact being used, and operant processes being involved in making the likelihood of a Mand being used increase (Ibid p. 197). A difficulty with this conception is that internal to theorising about the origins of Verbal Behaviour there is a consensus that Verbal Behaviour first emerged about 100,000 years ago (Hayes and Sandford 2014 p. 114). Furthermore, there is compelling evidence that the capacity for a creature to learn through conditioning of any kind began about 520 million years ago (Ginsburg, S, & Jablonka, E. 2019 p. 293). So, if we think of things such as Cyanobacteria which our best theories tell us existed for billions of years before either the evolution of a capacity to be conditioned or a capacity to engage in Verbal Behaviour (Schirrmeister et al 2015 p. 777). Given the capacities Barnes-Holmes uses to illustrate behavioural functions we use to define an object, he appears to be committed to the view that Cyanobacteria did not exist until creatures with a sophisticated behavioural capacity emerged on the scene, with the capacity to engage with them is a manner sophisticated enough to develop a verbal and non-verbal repertoire in relation to them. It could be argued that this position is a behavioural version of Berkeleyan Subjective Idealism where we must deny the existence of a behavioural independent world.

            Barnes-Holmes would deny that he is engaging in any type of Idealism. He has argued he is not making some anti-realist argument, rather he is making an a-ontological claim about reality. He says that the anti-realist is arguing that either nothing exists beyond scientific language, or that scientific language doesn’t capture reality as it really is (Barnes-Holmes 2005 p. 68). Whereas he defines a-ontological claims as claims which remain silent on behaviour independent reality (Ibid p. 68). Furthermore, he would argue that Cyanobacteria do enter into the behavioural stream of our working scientists, so we are therefore justified in postulating their existence internal to our overall theory of the world. But given that our best theories tell us that entities existed prior to creatures who have behavioural capacities like the ones Barnes-Holmes mentions; why remain silent about their ontological status. Why remain happy to say that it is sometimes ok to talk as if these entities existed, or that to the extent that these entities enter a scientist’s behavioural stream we are justified in saying they exist?

            Why not accord these entities robust realist ontological status? The following passage is instructive as an explanation of Barnes-Holmes reluctance:

“Assumption 3, however, appears to preclude the possibility, in behavioural pragmatism, of finding a scientific truth statement that corresponds to an ontological reality. In effect, if the scientific activity of the behavioural pragmatist is the product of a behavioural history, then he or she can never claim to have found an ontological truth, because a different or more extended history may have produced a different truth (an ontological truth, by definition, is immutable, absolute, and final).” (Barnes-Holmes 2000 p. 198).

So, assumption 3 notes that even the behaviourist’s behaviour is a product of their learning history as they interact with their own behavioural stream. Which means that a different learning history would give them a different theory.  From this Barnes-Holmes concludes that since an ontological truth is by definition “immutable, absolute, and final”, contingent creatures such as us can never arrive at such ontological truths.

            But one wonders why he thinks that we need to parse ontological truths as truths which are immutable absolute and final. Ontology is the philosophical discipline which aims to discover the basic furniture of the universe. To discover what is in the most general sense possible. In traditional philosophy ontology was opposed to epistemology. Famously, Kant argued that reality as it is in itself is unknowable, and that we can only know reality in so far as it conforms to our mode of cognition. So true reality in its ontological form is something which we can never know. Hence, Kant argued that metaphysics should give up the proud name of ontology. In this sense of ontology, we cannot say anything about ontology; even statements about whether it is “immutable, absolute, and final” would be out of place because whatever the nature of ontology beyond our mode of cognition we would not be justified in speaking about it.  So, it is doubtful, whether Barnes-Holmes meant ontology in the sense the Kantian sense.

            Older versions of ontology which would have begun with Plato would involve studying entities in the world and trying to discover their accidental and essential natures. The discovery of essences would in this sense would typically be conceived of as “immutable, absolute and final”. Metaphysics in this sense would have ontological purport and is still studied to this day in philosophy departments. If Barnes-Holmes is critical of this type of ontology that is one thing, however, it doesn’t follow that because the behavioural scientist is critical of ontology in Plato’s sense that it should therefore become a-ontological.

            There is another strand of ontology which doesn’t involve appeals to ‘immutable, absolute and final” properties and that version of ontology was developed by Quine who Barnes-Holmes discussed in the paper. It is surprising therefore that he didn’t criticise Quine’s view of ontology directly.

            Barnes-Holmes argues that the issue of scientific truth is defined ultimately in terms of whether it achieves certain goals; and for the behaviourist the ultimate goal is prediction and control. So, ontology doesn’t come into the issue at all. He explicates his a-ontological position by arguing that “no fundamental or final or absolute assumptions are ever made about the nature or substance of behaviour independent reality” (Barnes Holmes 2005 p. 68). In effect they ignore issues in relation to realism-vs-anti-realism and stick doing behavioural science involving the prediction and control of organism’s understudy.

“Functional relations, at least in behavior analysis, are correlational, and no mentalistic, cognitive, or intentional act of reference from the response to an ontologically real stimulus is implied when functional-analytic terms are used in a behavioural explanation. For the behavioural pragmatist, therefore, a technical analysis of ontological talk will be cast in terms of patterns of stimulus-response-stimulus interactions, not semantic reference, literal meaning, or some form of word-referent correspondence. The procedural instruction “set the tone to between x and y cycles per second,” for example, could be interpreted as a relational network of derived stimulus relations (Barnes-Holmes, Hayes, & Dymond, 2001; Barnes-Holmes, O’Hora, et. al., 2001) or an instructional stimulus composed of Tacts, intraverbals, relational autoclitics, and the like (Skinner, 1957), or perhaps a combination of both interpretations (Barnes-Holmes, et. al., 2000). In neither case, however, is semantic reference or literal correspondence to an ontological reality included as part of the explanatory nomenclature. The technical terms of behavior analysis are simply empty with respect to ontological reality, and thus neither realism nor antirealism is implied.” (Barnes-Holmes 2005 pp. 73-74)

Despite above writing that the technical terms of behavioural analysis are ontologically empty, he notes that ontological talk is regularly used in report sections of daily articles and in ordinary scientific activities (ibid p. 74), arguing that such talk is fine as long as it doesn’t involve talking about the fundamental nature of reality.

            The idea that it is ok for behavioural analysts to talk about a behaviour independent reality when describing the results of their experiments, but this talk isn’t to be taken seriously as a description of the reality.  Part of the motivation is that Barnes-Holmes wants to avoid being committed to making assumptions about the fundamental or absolute nature of reality. So committed  is he to avoiding speaking about absolute reality is he that he goes as far to define something as true in so far as it achieves certain scientific goals (prediction and control). Aside from the fact that this assumption is arbitrary; why for example is prediction and control given priority over explanatory depth?

            More importantly the behaviourist can achieve their goals without bending over backwards and trying to eschew all talk about a mind independent reality. A more realistic goal would be to treat ontology as a part of science; and treat ontological commitment as the working out of what theoretical presuppositions cannot be done away with to make sense of our total theory of the world. If we cannot make sense of our experimental results without presupposing certain entities, then we are justified in admitting into our ontology. Quine, argued that a good technique to lay bare the ontological commitments of a science are to translate into the syntax of first order logic arguing that:

“a theory is committed to those and only those entities to which bound variables of the theory must be capable of referring in order that the affirmations made in the theory be true” (Quine, 1953 pp. 13-14)

And he used this technique to try and dis-entangle philosophical disputes on ontological commitment in subjects like mathematics.

            It could be argued that while mathematicians and physicists can do their work without solving debates about ontological commitment all behavioural analysts are asking for is the freedom to do their work without solving debates about ontological commitment. However, I don’t think that this is a fair way of interpreting the debate. The behaviourists aren’t just eschewing ontological talk in some innocent theory independent way; rather they are making truth relative to their own personal goals; and judging a theory as to whether meets those goals adequately. Anything outside of these personal goals are deemed irrelevant.

     Quine and the Pre-Established Harmony.

Barnes-Holmes describes the problem of homology as a problem which arises when we assume a correspondence between observation sentences and ontological reality (Barnes-Holmes 2000 p. 195). Quine’s attempted solution does indeed make assumptions about ontological reality. His assumptions are largely physicalist. He is assuming mind independent world with features such that an organism must track if it is to survive, he assuming that creatures whose neural structures with innate standards of perceptual similarity which tend to harmonize with trends in the environment will be more likely to survive than animals whose innate structures don’t harmonize with the environment (Quine 2008 p.204).  And this assumption is then used to explain how human’s shared neural perceptual similarity structures will explain how they converge on stimulus meanings of observation sentences.

            One may or may not agree with Quine’s account of how humans shared perceptual similarity standards lead to similar stimulus meanings which accounts for agreement on observation sentences. But his account is no more ontologically profligate than behavioural scientists appeal to group selection for cooperation to explain the emergence of frames of coordination in children’s ontogeny (Hayes 2014 p. 123). Or when behaviourists explain taste aversion in rats through natural selection resulting in the temporal parameters of classical conditioning becoming distorted (Wilson and Hayes 2018 p. 53).

            It could be replied that behaviourists like Barnes-Holmes have difficulties not with appeals to neural structures, or sensory receptors being impinged on. Rather they want such explanations cashed out functionally instead of in terms of shared structures (Barnes-Holmes 2000 p. 201). But this reading cannot be correct. Behaviourists emphasize the central importance prediction in their theories. And Quine’s appeal to shared structures is making predictions which are eminently testable by future science. It would be churlish to critique Quine for appealing to shared structures if this appeal involves empirical check points and predictions which could be tested. Furthermore, explanations of distortions in classical conditioning in taste aversion which behaviourists unproblematically study involve appeals to shared structure in rats.

            Barnes-Holmes though could argue that it is not just Quine’s appeals to shared structure which he has difficulties with. He also had difficulties with Quine’s appeals to Observation Sentences as part of his check points which are used to ensure that theories are tested for truth. Barnes-Holmes has difficulties with any non-pragmatic appeal to truth as his assumption 3 that scientists views develop as a result of their learning history means that with a different learning history, they may have held different theories of the world. Barnes-Holmes believes that assumption 3 means that we can never hold “immutable, final, absolute” ontological truths. But when Quine speaks about ontology, he is not speaking in terms of “absolute, immutable, and final” truths. Quine’s entire philosophy is built around the concept of radical revisability of our overall theories of the world, and no aspect of our web-of-belief is immune from potential revision; including mathematics or logic.  Therefore Barnes-Holmes criticisms completely miss the mark when it comes to Quine, as when he speaks about ontology he is never speaking about absolute, immutable and final truths.  

                         Bibliography

Barnes-Holmes, D. 2000. “Behavioural Pragmatism: No Place for Reality and Truth”, The Behaviour Analyst 23 pp. 191-202.

Barnes-Holmes, D. 2005. “Behavioural Pragmatism is A-Ontological, Not Anti-Realist: A Reply to Tonneau”, Behaviour and Philosophy 33 pp. 67-79.

Ginsburg, S., & Jablonka, E,. 2019. The Evolution of the Sensitive Soul.  MIT Press. Cambridge MA.

Hayes, S. 2014. “Cooperation Came First: Evolution and Human Cognition.” Journal of the Experimental Analys of Behaviour 101 pp. 112-129.

Kemp, G. 2017. “Quine, Publicity and Pre-Established Harmony”, Protosociology 34 pp. 59-72.  

Quine, W. 1953.  From a Logical Point of View. Harvard University Press. Cambridge MA.

Quine, W. 1974. The Roots of Reference. La Salle. Open Court Press.

Quine, W. 1995. From Stimulus to Science. Cambridge. Mass. Harvard University Press.

Quine, W. 1996. “Progress on Two Fronts”, The Journal of Philosophy. 93/4 pp. 159-163

Quine, W. 2008. “The Flowering of Thought in Language” pp. 478-484 in Follesdal & Quine (EDS) Quine: Confessions of a Confirmed Extensionalist.

Wilson, D, S, & Hayes, S. 2018 Evolution and Contextual Behavioural Science. Context Press. New Harbinger Publications Inc.

Schirrmeister BE, Gugger M, Donoghue PCJ. 2015. “Cyanobacteria and the Great Oxidation Event: evidence from genes and fossils”. Palaeontology 58: 769–785.

Skinner, B.F. 1974. About Behaviourism. Vintage Books. New York.

Introduction

            In this blog post I will discuss Noam Chomsky’s 1971 paper ‘Psychology and Ideology’ where Chomsky critiques Skinner’s popular science book ‘Beyond Freedom and Dignity’. In Beyond Freedom and Dignity Skinner claimed that freewill was an illusion and we could explain it away using an effective behavioural science. And once we did this we would be in a position to use behavioural science to engineer a more effective society than the one we currently live in.

I will argue that while Chomsky sometimes caricatures Skinner, and he is blind to the strengths of behaviourism as a discipline, his criticisms still hit the mark. Skinner’s remarks in ‘Beyond Freedom and Dignity’, exaggerated what the science of behaviour was capable of in 1971, and is still beyond what behavioural science can achieve 50 years later. To this end, Chomsky’s 1971 paper did the field of psychology and philosophy a favour with its terse criticism of Skinner’s attempt at popular science.

Psychology and Ideology

 At the beginning of ‘Psychology and Ideology’ Chomsky noted when reflecting on psychological claims that we need to ask two different kinds of questions: (1) What is the scientific status of the claims, (2) What social or ideological needs do they serve. He correctly noted that these two claims are logically independent. He argues that Skinner’s empirical claims are vacuous and completely without scientific merit. And he argues that because of their null scientific status they can serve the purposes of any would be dictator as rhetoric whether the dictator was on the left or the right.

            Chomsky even goes as far to argue that the these of ‘Beyond Freedom and Dignity’ results in the whole project being incoherent:

 “But if his thesis is true, then there is also no point in his having written the book or our reading it. For the only point could be to modify behavior, and behavior, according to the thesis, is entirely controlled by arrangement of reinforcers. Therefore, reading the book can modify behavior only if it is a reinforcer, that is, if reading the book increases the probability of the behavior which led to reading the book (assuming an appropriate state of deprivation).” (Psychology and Ideology: p.21).

This is a forced choice it is probable that there are some true claims in the book and some false claims in the book. The cause of us reading the book could be specified in various manners; being reinforced by reading that class of book in the past; a book recommendation from a friend whose recommendations in the past have been reinforcing. But Chomsky doesn’t ask what the cause of reading it is; he asks what the point of reading it. He says on Skinner’s central thesis, the only point of reading the book, is that it will modify behaviour. And it will only modify behaviour if it is a reinforcer, that is if reading the book increases the probability of the behaviour which led to reading the book (ibid p. 21). This is a strange interpretation of Skinner’s project. We have already discussed possible causes of reading the book. Possible consequences of reading the book are negative reinforcement; reading the book takes away boredom and increases the probability of reading more books like this in the future. Or the reader could be punished as a result of finding the book unintelligible, and excruciating to read, which will decrease the probability of reading books of this class in the future. Or the person could find the book positively reinforcing and this may lead to reading more books of this class and possible seeking a career in behavioural psychology.

            It is important to note that behavioural psychology doesn’t stand or fall based on what point each individual gets from reading one popular book by B.F. Skinner, there are now over a hundred years of behavioural research. And behavioural science should be evaluated on its own terms, which are the degree to which their principles have given us the ability to predict and control the behaviour of various organisms.

            Chomsky goes on to argue as follows:

“Consider the claim that reading the book might reinforce such behavior. Unfortunately, the claim is clearly false, if we use the term ‘reinforce’ with anything like its technical meaning. Recall that reading the book reinforces the desired behavior only if it is a consequence of the behavior; and obviously putting our fate in the hands of behavioural technologists is not behavior that led to (and hence can be reinforced by) reading Skinner’s book. Therefore, the claim can be true only if we deprive the term ‘reinforce’ of its technical meaning.” (Ibid p. 22)

Reading books in the past has been reinforced positively in our school and college environment. Doing so has led to reinforcing consequences in the past. Reading behavioural books in the past (Watson’s Psychology as the Behaviourist Views it), led Skinner to pursue a career in behaviourism, this has been reinforcing in terms of (discovering things which he finds reinforcing), and a long career (which he found reinforcing). Similarly, for us, the general public, reading in the past has been reinforced, if ‘Beyond Freedom and Dignity’ is in the stimulus class of popular science books people found reinforcing in the past then this may lead to people reading it. Lots of people reading it could influence behavioural change which will alter the probability of how people behave in the future. What courses they study, how they manage selection by consequences etc[1].

Chomsky’s critique though goes beyond the idea of reinforcing consequences of reading the Beyond, Freedom, and Dignity. He is also critical of Skinner’s casual manner of translating ordinary discourse into behaviourist language, and at the perceived lack of progress in behavioural science.

“Because of this unwillingness, there is also no discernible progress – today’s formulations in this domain are hardly different from those of 15 or 20 years ago – and no convincing refutation, for those who are untroubled by the fact that explanations can be invented on the spot, whatever the facts may be, within a system that is devoid of substance.” (Ibid pp 29-30)

Here it is fair to say Chomsky does have a point. Skinner had a penchant for inventing explanations for any behaviour or cognitive capacities in terms of reinforcement. And there was little indication that Skinner was overly concerned about testing the empirical validity of his claims about various complex behaviour being explicable in terms of reinforcement. Even today some behaviourists who have been heavily influenced by Skinner are critical of him for this tendency:

Evolution was for many years dramatically gene-centric…ontogenetic evolution was virtually ignored…behaviour analysis seemed to have made the opposite error…A good example is provided by the transcript of the recorded interview between B.F. Skinner and E.O. Wilson, in which almost every specialized, evolutionarily established behaviour put forward by Wilson was promptly interpreted by in Operant terms (Hayes and Sanford 2014 p. 115).

So, Chomsky’s criticism of Skinner casually translating every complex trait into something explicable in terms of operant conditioning is to the point. However, his criticism about a lack of advancement in behavioural science doesn’t stand up to critical scrutiny. Chomsky wrote ‘Ideology and Psychology’ in 1971 and his claim that there hadn’t been a huge advancement in behavioural science since around 1950 isn’t true.

             Behavioural science since 1950 had undergone rapid changes. Breland and Breland’s (1963) work on animal training demonstrated that instinctual drift would mean that non-human animals’ behaviour wasn’t as malleable as earlier naïve behaviourists thought. Skinner, who had long stressed both phylogenetic and ontogenetic factors playing a role in animal behaviour welcomed the work of the Breland’s. On Skinner’s way of thinking it was the behaviourists job to discover the different ways behaviour could be shaped and controlled through different schedules of reinforcement. The behaviourist wasn’t in the game of stipulating how malleable different organisms were. Nonetheless, despite the Breland’s work being congenial to Skinner’s behaviourism, for the public, instinctive drift made the thoughts of behaviourists gaining control over people and shaping their behaviour seem less threatening.

             As behavioural scientists continued to study human’s operating under schedules of reinforcement there was more reason to think that humans couldn’t be just shaped at a whim through schedules of reinforcement. Hundreds of behavioural studies on rule governed behaviour[2], have demonstrated that when humans were operating under rules this made them less sensitive to the contingencies of reinforcement. Children below the age of 5 could be shaped under schedules of reinforcement in a similar way to a rat, but once they passed 5 and could follow rules their behaviour became less sensitive to the contingencies of reinforcement (Bentall).

            Like the Breland’s work, work on rule-following sprung up from within behaviourism (Skinner 1963) and demonstrated that human behaviour was more complex than the behaviour of other organisms. The year Skinner wrote ‘Ideology and Psychology Sidman (1971) experimentally demonstrated that humans could derive untrained stimulus equivalence. And in the years since Steven Hayes (1989) demonstrated that humans could derive untrained relational frames (coordination, comparison, hierarchy, etc). The human under behavioural science began to closer resemble the human as described by cognitive scientists (built with innate constraints, have a species-specific capacity for productive reasoning etc), than it did the human as described by early behaviourists.  

Now Chomsky would parse some of these studies as resulting in the death of behaviourism as derived by Skinner.  However, things don’t have to be parsed in this manner. One could look at the work of behaviourists, such as Rescorla, Breland, Sidman, Lowe et all’s experimental refutations of previously held beliefs by behaviourists as a sign of an evolving healthy discipline.

One thing that should be emphasised was that when Skinner was writing ‘Beyond Freedom and Dignity’ work on rule following in behavioural science was being done which was already demonstrating that we couldn’t simply reinforce behaviours we wanted repeated, when people operated under verbal rules, they were less sensitive to the contingencies of reinforcement than non-verbal animals. Thus, even for behaviourists at the time Skinner wrote ‘Beyond Freedom and Dignity his work was outdated.  Society couldn’t be shaped in the manner Skinner wanted.

Chomsky goes on to quote Skinner’s claims about techniques we could use to control speech. Chomsky notes, correctly that Skinner’s science isn’t up to the task of doing such job. But he notes it would be abhorrent if such controls could be put in place:

Or consider freedom of speech. Skinner’s approach suggests that control of speech by direct punishment should be avoided, but that it is quite appropriate for speech to be controlled, say, by restricting good jobs to people who say what is approved by the designer of the culture. In accordance with Skinner’s ideas, these would be no violation of ‘academic freedom’ if promotions were granted only to those who conform, in their speech and writings, to the rules of the culture, though it would be wrong to go farther and punish those who deviate by saying what they believe to be true. Such deviants will simply remain in a state of deprivation. In fact, by giving people strict rules to follow, so that they know just what to say to be ‘reinforced’ by promotion, we will be ‘making the world safer’ and thus achieving the ends of behavioural technology (74,81). The literature of freedom would, quite properly, reject and abhor such controls. (Ibid pp. 30-31).

What Chomsky doesn’t note though is that Skinner’s philosophy always had safeguards in place so those under control had a means to resist any science of behavioural engineering.

            Since 1953 Skinner had written about countercontrol as a way organisms had of resisting being controlled by others. Spencer et al (2022) define countercontrol as follows:

“Countercontrol is a Skinnerian Operant concept that posits that an individual’s attempts to exert control over another person’s behaviour may evoke a countercontrolling response from the person being controlled that functions to avoid or escape potentially aversive conditions generated by the controller.” (Spencer et al p. 457)

Skinner had targeted our notion that people are free. He argued that people only described themselves as free when they could not identify the variables which were controlling their behaviour. He also noted that when people are under the control of positive reinforcers, they often describe their behaviour as freely chosen. He gave the example of state lottery which works as an implicit tax on people and noted that people think they freely chose to do the lottery. He emphasised the point that people value freedom because it is controlled by positive reinforcement and as a result does not occasion countercontrol (Delprato 2002 p. 195).

            Skinner warned that evidence of lack of countercontrol as an indication of “freedom” was dangerous. Belief that we are freely acting can lead to inadvertently being subject to long term aversive consequences resulting from our behaviour. This can happen when the controller is aware of these long-term consequences and the controlee is the ultimate loser in this scenario (ibid p. 195).

            Skinner argues that the correct solution to control is not to abolish it (he thinks this is impossible), but to analyse it and see if this is the type of control and consequence we want and if not to figure out a different type of controlling system to work within:

“Humans need (a) to eliminate aversive control (often a practical impossibility), (b) to identify positive reinforcement and other inconspicuous forms of control that have deferred aversive consequences, and (c) to substitute positive reinforcement contingencies without such consequences.” (Delprato p. 196).

Delprato argues that the above sequence is practically impossible, so we are in effect stuck with the use of countercontrol.

            As we discussed above since the mid-sixties, we knew that once people begin to follow verbal rules their behaviour isn’t shaped by the contingencies of reinforcement in the same way as non-human behaviour. Nonetheless, control and counter control are still facts in any society we live in. Relational Frame Theorists, behaviourists who have been studying emergent properties of verbal behaviour, have recently tried to tie countercontrol in with derived relational responding to see if we could use the concept now that our understanding of rule-following behaviour has expanded beyond Skinner’s conception (Spencer et al 2022).

            As things stand 50 years after Skinner wrote his ‘Beyond Freedom and Dignity’ we still are nowhere near developing the capacity to behaviourally engineer our society. While Chomsky sometimes caricatured Skinner’s behaviourist position. He was surely correct to note that Skinner’s proclamations about the capacities of behavioural science to engineer our culture went well beyond anything possible in 1971 or anything possible today.

[1] I am not hear arguing that we can explain people reading books entirely in terms of reinforcement, I am merely demonstrating that Chomsky’s quick argument for incoherence doesn’t work.

[2] See for example: Weiner et al (1964), Lippman and Myer (1967), Lowe et al (1983), Hayes et al (1986).

Childers et al 2023 argue that the debate between Chomsky and Quine and Skinner is revisited by contemporary debates on Large Language Models and Chomskian linguistics. They go on to argue that proponents of the view that data driven large language models mirror human natural language make the same mistakes that Chomsky critiqued people like Quine and Skinner ago 50 years ago. In this blog post I will argue that Childers et al largely misinterpret Skinner and Quine’s project and hence any connections they draw with Large Language Models and Connectionism are problematic.

            Childers et al argue that Quine and Skinner’s empiricism has collapsed under criticisms from Chomsky. And that Quine responded to this collapse by modifying his position his empiricism to accommodate Chomsky’s position by appealing to innate mechanisms.  They note that once such an appeal accepted then we are in a place where we are no longer empiricists, and our position is closer to that of rationalists. I would argue that this is an idiosyncratic reading of empiricism. Empiricists, like Hume and Locke, appealed to innate mechanisms explain how we acquired our knowledge of the world. It is just that the innate mechanisms they appealed to wouldn’t be sufficient to account for the complexity of human language and cognition. Quine’s externalized empiricism is of a piece with Hume’s except for in Quine’s case he is advocating for the innate mechanisms to be determined experimentally. Childers et al call this hybridised empiricism and note that it is empiricism only in name. It is unclear to me why Quine arguing that we determine what innate structures are necessary based on behavioural tests should be considered anything other than ordinary empiricism externalized.

            They do go on to make the further point that Quine’s speculations about the innate principles necessary are extremely vague. I would agree with them on this point; when Quine talks about analogical synthesis being the method in which we connect sentences with sentences he is extremely vague. Gibson (1987) wrote about Quine’s postulation of analogical synthesis being a postulated innate structure to be mapped by future scientists. In the 35 years since Gibson wrote those words there has been very little work done by philosophers influenced by Quine filling in the details of this project.

            There has though been scientific research into analogy from a scientific perspective in both behavioural science and cognitive science. Relational Frame Theorists are behavioural scientists whose experimental work on language has left them to abandon some of Skinner’s account of language. Thus, they argue their research shows that rule following in language changes how people respond to schedules of reinforcement, and they argue that human specific emergent properties such as the ability to derive stimulus equivalence, and relations of coordination, hierarchy, etc. These relational frames under contextual control exhibit the property of Mutual entailment, combinatorial entailment, and transfer of function (Relational Frame Theory an Overview p.62). In Relational Frame Theory they demonstrated empirically that analogy is a relational frame which typically emerges at about 5 years of age, defining analogy as the capacity to derive sameness amongst equivalence relations or equivalence-equivalence responding.  They think that these derived relational frames may play a role in linguistic productivity. However, it would be unfair to argue that they vindicate Quine’s concept of analogical synthesis as Quine’s vague formulations played no role in the experimental paradigm. And furthermore, it is impossible to tell the degree to which RFT is consistent with Quine as his formulations were too vague to match on to them.

            In cognitive science analogical reasoning has been researched in detail see for example Hofstadter (2013), and Gentre and Hoyos (2017). Gentre has being doing experimental work on analogies for over 40 years. Based on her experimental work she argues that children start deriving analogies from 7months of age. The discrepancy between her chronology of when analogies are acquired and the timeline in RFT can be accounted for in the lower criteria set for what counts as an analogy for Gentre. Gentre argues that analogies involve transfer of knowledge from one area to another whereas RFT theorists argue that stimulus-stimulus equivalence is necessary for something to count as analogy.

            In their ‘Analogy and Abstraction’ Gentner and Hoyos note a difficulty in explaining children’s acquiring of abstract analogies and their difficulty is like the one which faces Quine. We know from the psychological literature that children (1) prefer to extend their concepts based on bare perceptual similarities, (2) this results in them making concrete analogies, (3) children can extend their comparison classes through multiple exemplars.

Two of the primary ways of making an analogy are through projective alignment and through mutual alignment. Projective alignment is when people use an already well understood domain to illuminate another domain. However, when it comes to young children, they do not have a large store of well understood concepts, so it is difficult for them to use already understood concepts to explain a different concept. Therefore, young children typically use mutual alignment in their analogical abstraction. In mutual alignment analogies one discovers commonalities which were not obvious in either analog (ibid p. 3). Mutual alignment involves establishing a structural alignment between two-representations based on matching relations between analogues (ibid p.4).

The difficulty is to explain how children can go from using concrete analogies based on bare perceptual similarities to more abstract relational concepts. One problem is that presenting young children with exemplars which are not perceptually similar will not be helpful as the children won’t have the capacity to mutually align the two analogs. So, we are left with a mystery as to how young children form more abstract relational analogs. Gentner and Hoyos argue that we overcome this barrier through a process called progressive alignment. Experimental studies have shown that if young children are presented with abstract relational analogues they cannot pick up on the relationship. However, if they are first trained on concrete analogues and then later retested on the abstract relational analogues their performance improves dramatically.

This progressive alignment gives young children the capacity to move beyond bare similarity and acquire more abstract relations. Like in the case of RFT it is possible to use these experimental works to fill out Quine’s speculations, however, given the vagueness of Quine’s speculations it would be a stretch to call this work Quinean. Thus, I would agree with Childers et al that Quine’s notion of similarity and analogy is too vague to do the work he set for himself.

When it comes to Skinner, Childers et al err two major ways. Firstly, they argue that Skinner’s Verbal Behaviour book would be unheard of today if it wasn’t connected in people’s mind with Quine’s empiricist model. However, Skinner’s Verbal Behaviour has in fact spawned a massive experimental literature with hundreds of experimental studies into Skinner’s Verbal Operants (Sauter and Leblanc 2006, Jennings et al 2021). At first research inspired by Skinner focused on simpler verbal operants such as Mands and Tacts with a lot of research being applied research with people with developmental disabilities (Ochs and Dixon 1989). But over the last 15 years there has been an explosion of research into more complex verbal operants such as the intraverbal, and there is a massive increase in non-applied experiments and studies of people without an intellectual disability or autism (over 50% of the people studied have no diagnosis) (Jennings et al 2021). As the years go by the pace of experimental tests into Skinner’s Verbal Operants are rapidly increasing (Aguirre et al 2016).  There is no evidence, or reason to think, that any of this research was inspired by Skinner’s name being associated with Quine. In fact Quine’s name is virtually never cited in papers on Verbal Behaviour, or in papers about behavioural off shoots from Verbal Behaviour.

Secondly, Kenneth McCorquodale (1970), replied to Chomsky’s review of Skinner’s Verbal Behaviour and noted that when Chomsky criticized Skinner’s notion of the probability of a Tact occurring, he confused the probability of a tact being used in a particular moment, versus the probability of a Tact being used at any point in a person’s life McCorquodale (1970). Childers et al criticised the use of the notion of “momentary probability” as an idiosyncratic use of probability and went on to make the following point:

. “More importantly, his claim that under certain circumstances the relevant probability becomes “extremely high” is unwarranted, unless we already know how the language functions.” (Childers et al p. 221).

It is unclear precisely what Childers et al mean by “under certain circumstances”. One circumstance could be in studies on intraverbal acquisition. If a child is being taught to use an intraverbal, using multi-exemplar training involving the frame “the wheels on the bus go_”, what would the probability of a child saying “round and round”? Are Childers et al seriously suggesting that we cannot estimate the probability cannot be established in these circumstances? And given that we have established developmental milestones on things like stimulus equivalence and the training procedures used to elicit them are Childers et al seriously suggesting that we cannot assign probabilities in these experimental settings? Granted Chomsky’s point still stands, it is probably not possible to assign probabilities to certain words being spoken as people interact daily, but has little bearing on experimental control in precise settings which is what Skinner and Relational Frame Theorists were interested in.

            Ultimately what Childers et al argue is that both Quine and Skinner had to do in response to Chomsky’s criticisms is to progressively modify their positions with more and more innate architecture. There is little evidence to support this interpretation of Quine, see Gibson (1987), for a detailed exposition of the extended debate between Quine and Chomsky on innate architecture. However, I largely agree with Childers et al that Quine’s vague sketch of analogical synthesis wasn’t detailed enough or clear enough to account for our linguistic development. This is important because Quine was committed to giving a mechanistic explanation for our behavioural capacities. He argued consistently throughout his career that it is ultimately at the neurological level we should be looking for our explanations, that behaviour was just data to point us towards underlying mechanisms. If the data turned out to support a Chomskian type architecture I don’t think it would have much difference to Quine’s overall project of naturalizing epistemology.

            Skinner on the other hand was interested in discovering behavioural regularities. He was interested in underlying neuroscientific explanations only insofar as they helped in functional control of the organism in particular circumstances. Whether people are impressed with the literature inspired by Skinner’s verbal behaviour and the predictions it makes is one thing. But this literature needs to be engaged with we cannot stipulate a priori how much experimental control has been gained in a particular experimental setting. As things stand there is little evidence that empirical evidence that Skinner’s account will ever be able to handle linguistic productivity[1]. And this will make it extremely difficult to ever gain predictive control over more complex linguistic behaviour beyond Tacts, Mands and simple Intraverbals. Behaviourists inspired by Skinner such as Relational Frame Theorists claim that they can handle linguistic productivity but since there is little engagement with linguistic data it is hard to test these claims. Ultimately these behaviourist studies are engaged in tests of how the organism behaves not stipulating the nature of the organism’s brain. The historical story Childers et al tell of Quine, Skinner and Chomsky’s dispute being mirrored in contemporary debates on connectionism, Large Language Models, and standard computational theories. The work of Quine and Skinner are very different from each other and both of their views have very little in common with work ongoing in Artificial Intelligence.

[1] See David C Palmer 2023 ‘Towards a Behavioural Interpretation of English Grammar’ for a recent Skinnerian inspired attempt at understand grammar behaviourally.

The Extended Mind is a thesis by Andy Clark and David Chalmers which states that the Mind Extends beyond the brain and encompasses aspects of the physical world. They give an example of a person who has dementia who keeps a diary to remind him of things he needs to do, such as when to take medication, where things are stored etc. If the person with dementia has reliable access to this diary most of the time, then they argue that the information in the diary is part of his extended mind.

            In ordinary circumstances if I want to remember when to do something it is because the information is stored in my brain, and I can access to information to make decisions. I don’t have to always have access to the information sometimes I may forget, but in general I have reliable access to the information. Chalmers and Clark argue that it is arbitrary to consider the information stored in a brain which you can reliably access to be part of your mind but to think that information in your diary which you can reliably access isn’t.

The thesis is counter intuitive. And some people reject it because of this counterintuitive feel, arguing that the thesis extends our ordinary concept of cognition too far. However, this counter argument has little force. There is little reason to assume that our theoretical understanding of a particular phenomena should be intuitive at first. Logical coherence should be the test of the theory not whether it chimes with your folk-psychological concepts.

The argument of Chalmers and Clark focused on information within a diary, but today with our phones which we carry everywhere with us storing so much information, the argument becomes even more radical, with it implying that aspects of the internet that we can reliably access are part of our extended mind.

Intellectual Disability

Psychologist J.J Gibson wrote about affordances which are relational aspects of our environment which we could interact with. Affordances relate not just to features of the environment but to the suitability of the environment to an observer or agent. Thus, steep stairs are an affordance for a person who can walk but to a person in a wheelchair they are not an affordance. Our natural and social environments contain affordances for some people but not for all. What are affordances are depends on the intentions, capabilities, and interests of the individual. In general, we tend to build our environment in such a way that can help people access the affordances they need, for example, building wheelchair ramps.

People with an intellectual disability who are non-verbal have interests and desires, but as a result of being non-verbal they will have difficulties in accessing various affordances in their environment. It is for this reason that there are practices and regulations in place which ensure that organisations who care for people with an intellectual disability do everything possible to ensure that they facilitate their communication capacities. Doing this involves giving them access to Speech and language therapists, occupational therapists, and a full Multi-Disciplinary Team. Furthermore, staff working with them must be trained in things like techniques to facilitate non-verbal people communicating their needs. Various types of augmentative communication devices are used on the recommendations of SLT, and things are in place such as visual schedules and Picture Exchange Communication System (an augmentative system based on Skinners Verbal Behaviour).

We saw above when discussing the patient with dementia who had reliable access to his diary that this could be considered a part of his extended mind. In the case of a non-verbal person in a service, there are affordances in their environment they may want to access such as going for coffee, visiting friends, going for a bus drive etc. If they can use PECS cards to indicate what they want, and those cards are not reliably available they you are taking away a part of their extended mind. It would be analogous to taking away a person’s prosthetic leg. A prosthetic leg may be artificial, but it is still a part of the persons way of accessing affordances in their environment and the same thing would apply to augmentative communicative systems.

I would argue that a similar thing is true of staff working with non-verbal people in a service. The staffing team ends up as part of the extended mind of the non-verbal person they are supporting. Just like the diary of the person with dementia contains information he can access, and my I-phone contains information I can access, staff working with non-verbal people with an intellectual disability contain information and affordances which a non-verbal person cannot access on their own.

As discussed above, not providing a person with reliable access to their prosthetic leg, denies them access to affordances in their environment, and likewise not providing a non-verbal person with experienced staff who know them and are versed in communication training is denying them affordances in their environment. It is for this reason that there is such a massive push in policies to ensure that effective communication training is available, consistent staffing are maintained etc. However, in a lot of the literature this is spelled out in atomistic terms. Thinking of these issues in terms of the extended mind helps people think  more relationally, and emphasises how our environment, social network, and social supports are partially constitutive of our own minds.

In this blogpost I will consider Quine’s relation to two of B.F. Skinner’s Verbal Operants: the Mand, and the Tact. I will argue that Quine’s Observation Sentence is analogous to Skinner’s Tact, but that Quine makes very little use of the notion of a Mand. A Mand is one of the most studied verbal operants, and it is the first verbal operants that is targeted by behavioural scientists if a child is experiencing language delays. Yet Quine gives the Mand a very small role in his overall theory of language acquisition. I will discuss Quine’s reasons for not considering things like Manding as important in sketching his naturalized epistemology and then discuss the degree to which these verbal operants are separable and how their separability will affect Quine’s choice of downplaying of the role of Mand’s in his story of how we go from stimulus-to science.

Quine’s Relation to Empirical Psychology

While Quine didn’t use the same vocabulary as Skinner some of his concepts map effortlessly onto Skinner’s concepts. Thus, Quine’s notion of an Observation Sentence is the same as Skinner’s notion of a Tact (both are shaped by discriminative stimulus and social feedback). Nonetheless despite arguing for this functional independence between different Verbal Operants Skinner noted that for ordinary speakers the two may be entwined. Thus, if a person has acquired a label as a Tact the chances are that they will be able to use it as a Mand. Skinner gives four reasons to support his argument for interdependence. (1) Tact emergence may be facilitated by the acquisition of a Mand in the presence of the Manded stimulus. (2) The similarity between the stimulus that evokes a Tact and that that evokes a Man may be similar enough to affect a transfer. (3) Transfer may occur if care givers reinforce one operant as if it were the other, (4) Children early in life may acquire generalized verbal skills which result in both the Mand, and the Tact being acquired (Petursdottir et all p.60). Skinner was speculating about these matters but in the 60 years since he wrote ‘Verbal Behaviour’ their partial interdependence has been confirmed.

Quine showed no interest in Mand’s and hence he had little interest in how Mand’s and Tacts related to each other and affected the process of acquiring a language.  He was clear that because of his interest in ontology and epistemology he was giving an idealized conception of how we acquire language. He was interested in observation sentences because they are our entering wedge into language and hence to our theory of the world. He is explicit that the story he gives of how we go from stimulus to science is meant to be an impressionistic one. He fully acknowledges that his story may deviate from the story told by a fully worked out science.

            Quine’s position on this matter is dubious. He was critical of Carnap for engaging in make belief in place of making use of current scientific psychology.

“But why all this creative reconstruction, all this make belief? The stimulation of his sensory receptors is all anybody has to go on, ultimately, in arriving at his picture of the world. Why not see how this reconstruction really proceeds? Why not settle for empirical psychology?” (Epistemology Naturalised p. 74)

 Yet Quine himself is in effect engaged in a made-up story about how we go from stimulus to science. He justifies this as follows:

“Much of what is earliest and most urgent in language learning, furthermore, is a matter of neither stating nor assenting nor acting upon statements, but of importuning…But statement learning is what is relevant to our study, which aims at understanding the acquisition of scientific theory.” (The Roots of Reference p. 46)

“Anyway, I am not bent even upon a factual account of the learning of English, welcome though it would be. My concern is with the essential psychogenesis of reference would be fulfilled in fair measure with a plausible account of how one might proceed from infancy step by step to a logically regimented language of science, even bypassing English” (Ibid p. 92)

 Because of Quine’s emphasis on cognitive language, he is ignoring the messier pragmatic aspects of learning a language that are described by people such as the later Wittgenstein, and Skinner. There is a sense in which we can justify this; it is after all standard practice in the sciences to engage in idealizations. Quine could be parsed as sketching a scientific story of how we go from stimulus to science; we can use this abstract sketch and fill in the details as we learn more about the acquisition process. Nonetheless, Quine does seem to be guilty of holding Carnap to higher standards than he holds himself to.

Because Quine is only interested in descriptive language and its role in us acquiring our theory of the world; he claims he doesn’t need to think about things such as importuning or as Skinner would it Manding.  Skinner had noted that Mands such as asking for water would be controlled by an establishment operation of a deprivation such as thirst and subsequent reinforcement of the thirst being relieved, while the Tact for Water could be controlled by a non-verbal discriminative stimulus, which the person was reinforced for saying the word in the presence of. Thus, you would have two different operants controlling the one sound for ‘Water’. As we saw above Skinner thought that these two operants would in practice end up entwined. If indeed Mands and Tacts are intertwined, then this would affect any proposed psychologically realistic story of how we go from stimulus to science.

Lamarre and Holand (1985) did a study on children tested the independence Tacts and Mands with preschool children. They trained children up on the relations “on the left” and “on the right” as both Mands and as Tacts. The study found that when the children learned them as Tacts, they couldn’t generalize them to Mands without training, while when they learned them as Mand’s they couldn’t generalize them to Tacts without training. This indicated that the two terms were for young children at first functionally independent and that transfer from one to the other wasn’t automatic.

Lamarre and Holland’s original study has been criticised by Wallace et al 2006. They noted that Lamarre and Holland’s establishing operations may not have been clear. There is no indication whether the items the child is manding for or the reinforcement they are receiving for Manding are items the child desires. When this was controlled for in other experiments the transfer from Tacts to Mands occurred for preferred items. Wallace et al claim their study demonstrates how responses taught as Tacts can facilitate the establishment of Mands for high preference items. And they noted that their experiments showed a difficulty with Tact-to-Mand transfer for low preference items (in line with Lamarre and Hollands study). Demonstrating that their lack of transfer was more than likely caused by not using sufficiently motivating reinforcers.

Gamba et al (2016) have done a meta-analysis on studies into whether functional independence in Mand-Tact independence has been demonstrate empirically. And they noted that there has been 28 empirical studies into the functional independence of the Mand and the Tact since Lamarre and Holland’s original study. They noted that there have been 13 studies which have demonstrated the functional independence of Mands and Tacts, but that in studies which the stimuli tacted and Manded were preferred items transfer of function occurred (Gamba et al p.27). Whether these studies are sufficient to cast doubt on the functional independence of the Mand and Tact is hard open to interpretation. Gamba et al note that in some of these experiments the Manded items were present which may have served to evoke previous Tacts and this may need to be controlled for in future experiments (ibid p. 31).

Thus far the experimental results are not sufficient to conclusively demonstrate the functional independence of Mand’s and Tacts. Quine appears to have been agnostic on this issue. However, one questions whether he has a right to this agnosticism. He was critical of Carnap for engaging in make belief in his epistemology, but his own Naturalized Epistemology engages in as much make belief. Quine’s focus on observation sentences at the expense of things like Mand’s serves to distort our picture of how we acquire our language and scientific heritage. In abstracting away from these details Quine is giving a hyper-intellectual fantasy of how we acquire language he is letting the metaphor of the scientist being a disinterested theorist keying observation sentences to stimuli, blind him to the more pragmatic aspects of language acquisition.

In this blogpost I discussed Quine’s relation to Skinner’s Verbal Operants of Mands and Tacts. In my next post I will focus on linking Quine’s concept of association of sentences to sentences with recent empirical work on Skinner’s Verbal Operant the Intraverbal.