Can Reasons Be Causes?
How mental contents causally account for a subset of human behaviours
In this short digest, I’m going to vindicate the following claim: the contents of our thoughts, whether formulated reflectively, in dialogue with others, or encoded in writing, are causally efficacious.
This is far from an obvious position to take. In fact, it is contested vigorously by philosophers and scientists alike, as evinced in a myriad of philosophical positions like epiphenomenalism, identity theory, behaviourism, etc and some sentiments in computer science and artificial intelligence research, where syntactical manipulation is thought to be a sufficient analogue for thought.
Epistemic Justification
In the Western tradition, to give reasons for a position one takes, for a belief, for a moral principle or otherwise normative claim, is tantamount to providing rational support for them. Reasons are the hallmark of rationality.
Reasons operate in the sphere of explicit statements. This is part and parcel of the philosophical tradition of proceeding through argumentation. Argumentation can be oral or written, but imposes constraints on the connection between reasons and conclusions, which are the statements they support.
So giving reasons is a far more nuanced and complicated matter than merely buttressing your beliefs by recourse to other beliefs, matters of fact, perceptual reports etc. The reasons we give must be connected in the right way with the statements we seek to justify.
If we accept argumentation as the hallmark of the rational, then right off the bat we also accept, more or less, that there are two chief ways of constructing arguments where the connection between the premises and the conclusion is not accidental. We can deduce the conclusion from the premises, or we can lend evidentiary support to the conclusion in a way that makes its truth more likely.
The first method is truth-preserving in the sense that, if a deduction meets the criteria for validity, the conclusion necessarily inherits a truth-value of True. Given that the premises are true, then the conclusion cannot be false; in fact, it would be impossible for it to be false — this is the standard of validity. In this sense, validity transmits truth, or preservers it.
The second method is ampliative, which is to say that it goes beyond the available evidence. When we induce or abduct from particulars to a general conclusion, we have gone beyond the particulars to abduct some general principle that holds in future instances. Therefore, induction and abduction (the latter, dubbed inference to the best explanation, is a subset of the former) are ampliative in the sense that, when (and if — Hume & Popper did not think so) such an inference is justified, it provides us with a rational basis for predicting future outcomes of domains of states of affairs.
That argumentation is the hallmark of rationality is not uncontested. In the philosophical branch of epistemology, what justifies our beliefs is vigorously debated. Not everyone agrees that justification is a matter of internal mental states of the thinking agent. While they may not necessarily deny the role of argumentation, they argue that, at the very least, it is not sufficient to justify the beliefs that thinking agents hold.
Yet, the question what justifies our beliefs equivocates on at least two, though possibly more, meanings. It could mean: what, independently of what we think, is the factual justificatory factor/s? Or it could mean, given that justification is a desirable standard, what are the rules we ought to follow to maximize the justificatory support of our beliefs? Both of these are parasitic on the following ambiguity: is justification a mental state that I have access to and can revise, or is it exhausted by an unconscious physical substrate that performs some cognitive role: memory, perception, learning, computation, etc.
This equivocation resolves, or rather appears to resolve, across the varieties of philosophical conceptions of justification we classify as internalist and externalist.
The internalist locates the locus of the desired connection between beliefs and truth within some reflective standard the rational agent must enact. By reflective standard I mean that the agent purportedly plays a conscious role in ensuring the net justified status of her or his web of belief. The externalist, by contrast, absolves or divests that rational agent from such a reflective standard or responsibility. The justificatory standard is met in lieu of the agent’s conscious effort, paradigmatically by a set of reliable belief-causing and transmitting processes.
What constitute reliable belief-causing and transmitting processes? By its very definition cognitive processes that are more likely to cause and transmit true beliefs. Perceptual states, memory, and valid forms of inference, to name a few, come to mind as candidates that satisfy this role.
It isn’t surprising then that some have viewed externalist positions as deflating altogether the concept of justification. For if justification is a matter of the reliable connection between belief formation and truth, then in many instances these appear to be carried out non-reflectively. To give the argument deeper roots, we can argue that the process of evolution by natural selection has equipped us with these reliable mechanisms, so in a sense our common and casual beliefs are broadly well-founded. This latter premise has a pragmatist tinge, since the status of these processes is contingent on their average outcomes. As long as the outcomes are truth-preserving, they qualify as justification-conferring processes. But this standard makes no requirement on the rational agent to know why these processes work; merely that they do seems to suffice.
Perhaps it has become clear to the reader that thus far no strong case has been made for these differing standards to be mutually exclusive. Why must our beliefs either require the onerous authorization of rational (reflective) justification, or wholesale concession to processes that operate without conscious effort? It seems that both reflective rationalization and third-person transmission methods play a crucial role in ensuring that, broadly speaking, our beliefs do not diverge from what is the case, and that these do not conflict internally with one another.
What I’ve just discussed so far may seem to be completely unrelated or divorced from the question I raised at the outset (whether reasons can be causes) and the position I have pledged to vindicate: that the contents of a thinking agent’s beliefs play a causal role in the world, chiefly in causing a subset of that agent’s behaviour, and that, all in all, belief-contents have causal efficacy.
Why raise this question at all, you ask? Doesn’t it seem trivial, or kind of a category mistake? I haste to affirm that the question is not just non-trivial, but pegs a whole array of intellectual conundrums that maintain relevance in our contemporary moment, and thread their way back through the Western intellectual tradition to its inception with the ancient Greeks.
To illustrate, take two of the chief varieties of internalist standards: foundationalism and coherentism. Within either of these conceptions, the rational agent must make good of either of these norms. In the foundationalist sense, she must make sure her beliefs reduce or trace their lineage back to certain basic beliefs, which are accorded special justificatory status. These beliefs are, perhaps, self justifying. Common candidates for basic beliefs include beliefs caused by perceptual experiences. Now as a rational agent, if she wants her non-basic beliefs to be justified she must ensure that whatever beliefs she acquires, these must be inferentially reducible to those self-justifying basic beliefs.
In the coherentist sense, she must ensure her beliefs form a perfectly coherent set or a set meeting some threshold of acceptable coherence, such as one that maximizes relations of inferential and explanatory support. Neither of these standards appears satisfiable without some conscious access to the source of justification.
Contrast this with externalist accounts. The externalist, broadly speaking, does not require that the agent justify her beliefs in the standard sense we have specified. Beliefs can be justified in lieu of rational reasons that the agent may give for them. This does not imply that an externalist eschews standards like foundations or coherence. It’s just that these standards may be satisfied whether or not the agent is aware of them. For example, imagine that there is some cognitive mechanism that weeds out contradictory beliefs whether or not the agent is aware of it. This cognitive mechanism, absent the agent’s reasons, would ensure that the agent’s beliefs cohere. This changes the meaning of justification from buttressing beliefs with good reasons to an external standard like reliability. If beliefs are caused by reliable truth transmitting and preserving processes, then we needn’t worry about justification in the traditional sense.
And yet, externalism cannot be the whole picture, or sufficient alone. It is precisely because we reflectively consider how we acquire knowledge that we, in turn, improve the standards for knowledge acquisition. If this weren’t the case, we would interact with nature just as other species do: through readymade processes that change or improve only through genetic and epigenetic transmission.
This ostensible tension between internalist and externalists accounts of epistemic justification is not normatively nor descriptively irresolvable. Both varieties of standards come into play in the transmission and maximization of truth, sometimes simultaneously and other times with one overriding the other. The very distinction, in fact, between these types of standards remains somewhat pragmatic: while rationality can make us see the blindspots of the cognitive mechanisms upon which we defer to in our day-to-day activities, known as cognitive biases, the very practice of rationality as a rarified and strenuous pursuit of assumptive explicitization, would not be possible without these largely automatic processes running in the background.
If you’re confused thus far, I want to give away the punchline here: reflective standards of justification spill into cultural and institutional modes of justification. If this weren’t the case, the evolution of the methodology of science, broadly connecting relevant evidence to consistent bodies of theory, would be inexplicable (I despair of making sense of epiphenomenonalist accounts). In other words, rationality (and empirical science as a subset thereof) as a collective project emerges from the codification of the rules of rational investigation, whether empirically of nature, or of the internal consistency of attending theory (mathematics included). The rules of rational investigation are a product of the thread of cultural practices that sanctioned and incentivized robust reflective standards. If this story is true, then we cannot divest reflective standards from the causal role they play in the broader reconfiguration of society and collectives, regardless of how indirect and attenuated that effect is in the broad scheme of things. Whatever physicalist story we tell of the constitution and character of the mental, it must be able to accommodate this narrative.
Mental Causation
Reflective and non-reflective standards of justification notwithstanding, we can see that the conversation here is confined to beliefs. In the philosophical jargon beliefs are intentional states that have a word-to-world direction of fit. However way we theorize the contents of beliefs, they aspire to depict, broadly, possible states of affairs and, narrowly, actual ones. Conversely, desires constitute intentional states that have a world-to-word direction of fit. They are intentional states of how we would like the world to be, such as satisfy our desires.
One conception of rationality construes reason as the instrumental pursuit of such preferred outcomes through attendant beliefs. If reason is subordinated to the maximization of our preferences, then reflective virtues of truth maximization acquire an instrumental context. It does not follow that our preferences mandate that we maximize the truth, only that the preferences we want to actualize are in some cases serviced by relevant accurate information or truth. On the other hand, not all preferences are serviced by truth; in fact a delicate equilibrium must be nurtured, many of whose attendant assumptions are in fact constructed rather than merely “described”. There are too many facts in the world for the relationship between goals and beliefs to be linear. Furthermore, the way those facts are “fixed” is heavily underdetermined since at least a subset of facts are cultural constructions, for example, like religious dogmas that undergird modes of social engagement. The chief point, nevertheless, stands: if rationality is the instrumental pursuit of preferences, these are affected by beliefs, many of which are adjudicated by a veridical correspondence to the world.
The upshot of the foregoing discussion is that if we try to account for human behaviour, we must then take into account that the human being nurses a complex representation of the world in her mind augmented infinitesimally by the representational affordances of language. Beliefs form a subset of mental states that have propositional contents. The propositional contents are, in all likelihood, a contingent element of the more basic mental attitude that forms assumptions and commits to those assumptions. How those assumptions are formed is a complicated matter. It is probable that we share cognitive rules for arriving at quick assumptions with proximate species.
Beliefs are a special subset of cognitive assumptions that are linguistically encoded. This is not say to that language is a precondition for belief. A large subset of beliefs are likely arrived without requirement for language at all. The scope of expectation from moment to moment, for example, relies on far more basic combinations of perception (perception construed broadly to include proprioception, interoception and the like), memory and mental modelling. However, when encoded linguistically, beliefs give granular expression to those non-linguistic assumptions via linguistic form. Because of the modularity of linguistic representation, the beliefs expressing such basic assumptions can be fine-tuned, expressed at varying levels of granularity, and become subject to revision in a manner far more flexible than non-linguistic representations. This is, in part, because the scope of fallibility increases exponentially with the representational capacities of language. Moreover, linguistically saturated beliefs extend beyond more basic non-linguistic assumptions because they come equipped with a matrix of publicly meaningful ontological commitments. By ontological commitments I mean the sorts of entities of which we implicitly or explicitly assume the world is composed. Beliefs blend with other forms of mental contents, such as imagistic modelling, to form a complex brew of internal representations that allow agents to contemplate possibilities and pursue outcomes according to their acquired preferences.
I have somewhat set the stage for how it is that mental contents affect behaviour. But before going further with my defence of mental causation, let me put forth some caveats that set the foregoing discussion in context with contemporary debate in analytic philosophy.
Since we’re invoking the notion of causality, it rewards to unpack first, just what the causal relata are (are they events, entities, facts or something else?), and second, just what the causal relation consists in (namely, just what does it mean to say that one thing causes another). Unsurprisingly, no agreement prevails about either of these questions, in spite of the centrality of causation to science.
I take causal relata to be spatiotemporal events, which is to say that they are immanent, i.e. they inhere in the events themselves. Further, causal relata individuate relative to a level of description within a larger body of theory. By level of description I mean the compositional scale that we describe some phenomenon or other. By body of theory, I mean the relevant general principles corroborated by evidence that can be brought to bear to explain a given phenomenon.
The individuation of causal relata is constrained by observation and discovered structural features of a given level of description. For example, at the social scale we individuate persons and institutions as causal agents, but at the biological scale, we individuate cells or systems of cells. Causal relata, therefore, must capture effectual nodes in nature relative to some corroborated description and are in this sense somewhat coarse-grained (meaning that conceptual joints correspond to natural kinds or properties intrinsic to mind-independent phenomena). The ascription of coarse-grained is contrasted with fine-grained, which takes individuation to be as fine as our linguistic posits allow. Those who hold the relata to be as fine-grained as, let’s say, propositions, hold the relata to be facts rather than events or some other candidate category. Since facts are not ordered spatiotemporally, these views characterize causal relata as transcendent rather than as immanent, and adduce ontological categories like objects to do the work in immanent relations, like pushing and shoving to take a simple example.
The granularity of causal relata is constrained, as I have already suggested, by a regimented combination of ordinary language perceptual terms and conceptual joints admitted by reigning theory in a given level of description. These conceptual joints are at best proxies for the underlying structures that yield tested and observed natural regularities or patterns.
Under this conception, causation is an epistemic schema heuristically wedded to our evolved cognitive context that functions as a proxy for the structure of ordered states of affairs. As such, it tracks transmissions of change against background variables held constant.
As regards the causal relation, there are three candidates: a) it is primitive, b) it is reducible to something more basic, and c) it can be eliminated altogether. Each of these views can be advanced on somewhat plausible grounds. The case for primitivity is that the reductive accounts do not altogether eliminate implicit reference to the notion. Causality is presupposed in reductive accounts that aim to reduce causation to process or probability, so they fail to properly analyze it. The case for elimination, as propounded by Russell and Quine, sees causation as a parochial, albeit useful, schema to be superseded by science. A problem with eliminativist accounts is that they ascribe ontological priority to mathematical equations, which lack reference to causes, events, or even objects. Yet substituting equations for reality, or relying on a lean ontological schema that admits only of mathematical objects and the most primitive of physical constituents (e.g. fields of force), can lead to conceptual absurdities that are best avoided. As regards reductionism, there are two chief candidates: a) probabilistic and b) process accounts.
Probabilistic accounts come in many varieties, prominent among them being David Lewis’s notion of counterfactual dependence, but they unify on characterizing the causal relation as essentially probability-raising: occurrence of the cause raises the probability of an effect. As regards the ultimate nature of causation, I will refrain from committing to either a probabilistic or deterministic account on grounds that I do not think that either notion, even in their most rigorous formulations, is wholly adequate on its own.
However, I will take aspects of the process account of causation as the theory that best captures the notion of causal relation. While many of the probabilistic accounts like counterfactual dependence do a great amount of work in elucidating the nature of the causal relation, the process accounts seem to metaphysically come closest to what the invocation of the relation either in ordinary language or the sciences aims to capture. In short, the process account holds that a causal process is a physical process that transmits a mark in a spatiotemporally continuous way. And a causal interaction “involves a spatiotemporal intersection between two causal processes which modifies the structure of both.” This account exhibits the virtue of weeding out pseudo-processes like casting a shadow.
The question then becomes, is the process account of causation compatible with mental causation? In what sense can beliefs and reasons for beliefs be said to constitute physical processes that transmit marks in a spatiotemporally continuous way? You might be thinking that the answer lies with neural activity in the human brain. Beliefs are physically realized by electro-chemical signals propagated across massively parallel and frequency-locking networks. Which causes our behaviour then: neural patterns or the contents of our beliefs? Are the contents of beliefs just neural patterns? Say that they are. How is one belief encoded differently from another belief in a neural pattern? Notwithstanding not being able to specify how neural patterns correlate to mental contents, there’s something far more bizarre going on here. We cannot observe nor infer beliefs by observing the neural patterns, unless we have sufficient requisite knowledge. Somehow, beliefs are only internally accessed, even though we might impute them to others from observed behaviour. Whatever the correlates to intentional states turn out to be, what matters is that the contents are accessible to the subject reflectively. This feature muddles the causal picture because it allows for loops of revision unlike the loops instantiated in computational systems: computer algorithms can modify themselves through feedback but cannot yet globally represent themselves the way we do through conscious thought.
The chief problem with the process account is that it is not clear how causal processes of the kind it identifies as paradigmatic scale up into macro processes or patterns that are, far more commonly, required to describe, explain and understand complex phenomena. Think of phenomena like Brownian motion or economies. Explanations of these processes rely usually on covering-law subsumption: that is to say, they are explained by being assimilated into a broader, either nomological (law-like) or statistical general patterns that do not prima facia make any commitment to causes.
As I see it, the resolution has to be along the following lines: causation as understood at the macro-level of observation and even first-person experience, ought to be, partially grounded in the fundamental interactions that we have so far discovered (i.e. the strong, electroweak, and gravitational forces). These fundamental interactions, in virtue of very specific physical properties like range, energy levels etc, yield the structural stability and baseline regularities that are inferable at our scale of observation.
However, while recourse to fundamentality helps explain aspects of the structures that we observe, it is not sufficient to explain higher level organizational features. For example, it is not by recourse to fundamentality that we explain speciation in the phylogenetic tree. More contingent information is required about environmental features to explain, for example, the evolution of vertebrates. Furthermore, the fundamental interactions are compatible with a wide range of causal behaviours at the macro-scale. In other words, there are causal layers that emerge in light of compossible (possible in conjunction) and/or contingent environmental factors at higher levels of organization.
This view hinges on making proper sense of the phenomenon of emergence, and whether the claim that novel causal powers emerge at higher levels of organization can be cogently advanced.
Getting completely precise on the notion of causation, however, is not altogether necessary for our argument since the notion of causation is ultimately up for revision given the evolving ontology of physics and how purportedly fundamental phenomena scale up to higher levels.
Why then bother at all with expounding on the notion of causality? Because the article aims to assimilate reasons and mental contents within a naturalistic account consistent with the fundamental and special sciences. This implies that we ought to be able to tell a story that specifies a mechanism consistent with the available evidence as to how the contents of thoughts induce effects in the world. I take mechanism to mean a causal account that hones in on relevant processes while not invoking processes inconsistent with corroborated theory and available evidence. In other words, reasons do not merely do work in a folk-psychological way, which can usually be explained away by refinement of the ontology via science (namely reduction of folk categories to relevant physicalistic correlates), but rather factor consequentially within objective causal relations without erosion to their ontological status.
Why go this route at all to vindicate the causal efficacy of reasons instead of pursuing some reductive route? Because we lack an explanation at a lower-level description with the predictive power that mental properties (mentalese — the hypothesized language of thought, among them) yield. This is not to say that such an explanation is not possible, but it is conceivable and, given the evidence, probable that the physical world permits configurations that instantiate intentional states. The scope of the project of vindicating the causal efficacy of beliefs and reasons requires naturalizing intentionality, namely, subsuming it into the physicalistic picture we subsume the rest of nature.
If we survey the evidence of the evolution of human societies, we would find it impossible to explain their configurations without identifying ideas and ideologies as causal factors. It may be possible that ideologies are completely epiphenomenal, and extra-ideological factors account exclusively for civilizations trajectories. But the latter would be the extraordinary claim, requiring extraordinary evidence, rather than the former, which is, in many ways, heavily confirmed. If the latter were the case, it would be accidental that Christian societies produced frescoes depicting Biblical stories, that the adoption of the Confucian ethic resulted in the stable and long-lasting Chinese Empire, that the prohibitions prescribed in Christianity, Judaism and Islam result in behaviour that observes those prohibitions. It appears much more plausible instead that the way we codify the world within belief-systems shapes individual and collective behaviour. While belief systems interact in complex ways with other factors like genetic inheritance, chance, geography, history etc, to deny their causal influence altogether would be tantamount to rendering inexplicable large swaths of the cultural evolution of humanity. The philosophically challenging task remains to demonstrate how the causal power of beliefs is indeed possible within a physicalistic view of nature.
More recently this conundrum has been formulated as the Exclusion Argument by Karen Bennett.
Bennett succinctly captures the tension between philosophical arguments for or against the causal efficacy of mental states by laying out five theses that seem jointly incompatible:
(a) Distinctness: Mental properties (and perhaps events) are distinct from physical properties (or events).
(b) Completeness: Every physical occurrence has a sufficient physical cause.
(c ) Efficacy: Mental events sometimes cause physical ones, and sometimes do so in virtue of their mental properties.
(d) Nonoverdetermination: The effects of mental cause are not systematically overdetermined; they are not on a par with the deaths of firing squad victims.
(e) Exclusion: No effect has more than one sufficient cause unless it is overdetermined.
Bennett’s denies (e) and argues for the joint truth of (a)-(d). By denying that the obtainability of more than one sufficient cause implies overdetermination, Bennett preserves the distinctness of the mental and physical, avoids overdetermination, and vindicates the causal efficacy of the mental, all without violating the completeness thesis: commitment to physicalism.
The intelligibility of each of these theses even in light of commitment to our best body of scientific theory is far from settled. More concisely, without some background agreement of an account of causality and a formulation of the physical, these theses are rather nebulous. For example, (b) completeness, is meaningless without a robust definition of physicalism, while with one, it’s completely vacuous, except by reference to non-physical causes. I don’t have a coherent notion of non-physical cause at hand, and I despair of making coherent sense of any such notion. My own definition of physical is anything consistent with our best theory of physics, namely any permissible configuration given the laws of physics, assuming the laws remain consistent with systematic observation. (d) Overdetermination, on the other hand, is a conceptual error. There are no genuine cases of overdetermination in nature. The illusion of overdetermination arises from the general terms of natural languages, which in turn enable us to conceptualize events in terms of types. There are no types in nature, only particulars. Every appearance of overdetermination can be reduced to a description that does away with such an illusion: every minute difference corresponds to a difference of effects, whether or not that disrupts the overall pattern or not. This is a version of the supervenience claim. And finally (e) subscribes to an acidity (number) of two causal relata. However, the idea that causal relata have an acidity at all is, in my view, an artifact of our conceptualizations, not a genuine feature of nature. So to say that no effect has more than one sufficient cause is utterly nonsensical. Effects and causes are simplifications that aim to isolate some stable relationship in nature; but that we could do so in a way that does not leave anything out does not, to my mind, appear plausible.
Therefore, as far as I’m concerned each of these statements are ill-construed. My only wish in this brief is to vindicate that our minds have contents and that these contents play a causal role in explaining our behaviour. In some sense this seems obvious; what doesn’t seem obvious is the causal story behind it. In isolating intentionality as the feature that confers this power, we have to specify how it is that intentional states stand in relation to their neural substrate in a way that differs from the way that software stands in relation to the hardware. The causal powers of the computer are entirely explicable, while the causal powers of the mind are as yet not.
Of course, the reductionist comes along and says that, if mental contents are identical to some physical substrate, then we might as well be parsimonious and ascribe causal efficacy to that physical substrate alone. This is where (a), Bennett’s thesis of the distinctness of the mental from the physical, can be invoked. But even if we reject distinctness, it seems plausible that even if mental contents were identical to some token physical states (rather contents are realized by the network as a whole), which I hold, their efficacy lies with their being accessible to the agent internally. Thus, even if we deny the distinctness of mental and physical properties, the locus of efficacy ought to be the physical network that realizes reflective awareness. Therein lies the causal efficacy of human intentionality. Mental causation of the type I’m vindicating ought to then be a byproduct of reflective awareness.
Reasons & Consciousness
The physicalistic imperative within analytic philosophy has yielded some strange views about the nature of the mind. Some species of intentionalism have sought to explain how a living organism realizes mental representations without consciousness. As if consciousness were something that you affix post fact after the functional and thereby causal mechanisms have been specified, kind of like universal suffrage on constitutionalism and accountable government. While consciousness is very much likely further analyzable (decomposable into more basic parts — this also being contentious — see panpsychism for alternative views) as a concept, its hazy referentiality points toward a property or set of properties that form the groundswell for any representation at all. I take consciousness, not uncontroversially, to be continuous with the emergent property of being alive, beginning with sentience of the body (the exact meaning of emergence is contested, so I only employ it here in the colloquial sense). Being alive is no more than a set of ongoing physical processes, but it is the summation of those processes that instantiate sentience as a feature for the maintenance of the organism in time.
We don’t yet know whether consciousness is a property that emerges past a critical threshold or a property that admits of gradations. However, since we have good reason to believe that some of its qualities can be found in simpler organisms, it’s not implausible that the totality of what we are acquainted with as conscious experience is the assemblage of many moving parts. We don’t yet know, however, how it is that simpler organisms with nervous systems that may plausibly lack first-person experience can then be scaffolded to more complex organisms like us that do.
Mental causation of the type I’m defending requires consciousness. Just so we’re clear: while philosophers disagree about how to analyze consciousness, many actually concur that what we term thoughts or propositional contents are not dependent on nor accompanied/saturated with affect or first-person experience. While this position is today less popular, the separation of dependence is still the orthodoxy in some corners. The motivation is largely the following: to say that thoughts do not require/depend on first-person experience enables their assimilation into a physicalistic-causal picture with less recalcitrance. Meanwhile, the idea that our third-person descriptions leave something out, i.e. are beset by an explanatory gap, creates a seemingly intractable problem for physicalistic naturalism.
Now, if only things were that simple. I counteract this narrative by averring that phenomenal consciousness including all the complexities of accompanying affect are continuous with cognition as such. While this position is defended by some camps, it’s also beset by the following problem: the vast majority of information processing is unconscious not in the Freudian sense of being repressed, but simply inaccessible to conscious thought. Do we then want to say that first-person experience is basic to or continuous with cognition, given that it’s probably not instantiated in most of life? This is where the gradation hypothesis may be leveraged to alleviate the apparent tension between third-person descriptions and the seemingly inexplicable and anomalous phenomenon of first-person experience. The gradation hypothesis, without further elaboration here, gestures to explain sensations as simply co-extensive with any organism’s environmental reactiveness. That is to say, environmental perturbations disseminate through the nervous system as electro-chemical signals that generate a global response at the somatic level: that global response, and this is where things get hazy, must be sufficiency integrated to be internally experienced.
Not only is conscious awareness (as an extension of affect) not cognitively separable from thought, but the contention is that the latter is ontologically dependent on the former: we can have thoughts of the kind we do because of ancillary conscious experience; the latter is the stage upon which thought unfolds. If this is true, phenomenal consciousness then is not some additional, causally inert ingredient that can be squeezed to conferring “what it’s like-ness” or qualitative experience to the rest of cognitive processes. Rather, experience and first-person qualitative states form the conditions of possibility for the kind of reflective standards we discussed earlier.
There remains an outstanding contention: is the supposition of phenomenal consciousness as an ancillary condition of intentionality relevant to the premise that beliefs and other intentional states are causally efficacious? If we can functionally replicate the causal powers of intentionality without the need for first-person awareness or phenomenal consciousness, what does this imply about the role of the latter? Do computing systems that model feedback like artificial neural networks through functions like forward and back-propagation sufficiently simulate the information-processing of humans to be able to instantiate their spectrum of behavioural outputs? If so, the internal states of ANNs may be thought of as analogues to human belief networks that then constrain/regulate/define their search-space and its preferential evolution within it. Hypothetically, an ANN with the same output capacities of a human must also instantiate the same causal powers: a functional parity must hold between the two. This question remains to be resolved.
