“Metaphysics is a restaurant where they give you a thirty thousand page menu, and no food.” (Pirsig, 1991, p.63). This quote is symptomatic of the criticisms levelled against metaphysics. Running with the metaphor, metaphysics appears to offer promising truths, yet one is commonly left undernourished and unsatisfied by metaphysical answers. Historically, many have seen this as indicative of a flaw within metaphysics itself. This essay shall be an attempt to show how the challenges can be used constructively to refine the discipline. Beginning with the anti-metaphysics of Neo-Humean logical positivism, moving on to Carnapian linguistics, I will lay out why these attacks fail to show metaphysics is impossible. The discussion will re-orientate around Kant’s conception of metaphysics, culminating in a stressing of the relevance of structural analysis in a science-informed, naturalised metaphysics.
First, an elucidation of terms will be given. By ‘impossible’ I mean there is nothing for metaphysics to understand, “in that there are no facts about or characteristics of empirical reality which can legitimately give rise to proper metaphysical enquiry” (Copleston, 1949-50, p.65). As will be shown immanently, ‘impossible’ in this sense in commonly conflated with ‘meaningless’ in verificationist criticism. I take the term ‘metaphysics’ to be an amalgamation of two approaches: ‘descriptive’ (analysing the categorical features of reality) and ‘revisionary’ (speculative challenging of aforementioned categorical schemas) (Copleston, 1949-50, p.65).
An ancestor of modern verificationism, Hume’s criterion of meaning postulated that proper inquiry could only contain matters of fact (a posteriori reasoning) and relations of ideas (analytic statements). Hume argued metaphysics could fulfil neither criteria, urging us to cast any such offending tome “... into the flames, as it can contain nothing but sophistry and illusion” (1748, §XII, Part III). Inspired by such rigorous empiricism, the logical positivists wrought the principle of verifiability: a sentence is meaningful provided there is a method for verifying it (Audi, 1999, p.739). There must be a way of discovering whether it is conclusively true or false; otherwise one is reduced to making meaningless pseudo-statements. Rather than attempting to reveal truths through speculative a priori foundations, supposedly leading to super-empirical nonsense, they undertook “a process of elimination, leading to discovery” (Ayer, 1971, p.45). This approach stands as a distinctive renunciation of transcendental metaphysics. Kant took up a similar challenge against revisionary metaphysics, but on epistemic grounds, whereas Ayer would see all of metaphysics dismissed through rules of semantic logic.
Two forms of the verification principle were propagated by Ayer, both are open to criticism. The strong form asserts a sentence is verifiable if and only if its truth can be conclusively established in experience. Such an absolute concept of meaning undermines itself, the principle cannot conform to its own criteria of meaningfulness; a foundation that excludes its own truth doesn’t seems to offer any real foundation at all. Its inflexible criteria lead to self-referential incoherence. However, Ayer saw the weak form as offering an escape from such incoherence. The weak form only requires it to be possible that experience could render a statement probable. Statements which aren’t relevant to any possible experience contain no factual content and are therefore meaningless. Yet this apparent reconstruction of the principle still fails to operate as an adequate criterion of meaningfulness. Suppose I state that “the world may eventually end in nuclear war”, this is a statement about the possibility of some future event, there is no way to verify the possibility of this happening. No experience can give us temporal access to the future in such a way as to render probable truth or falsity of a statement like this, because we consider it to be a loaded possibility populated by too many variables. As the statement is certainly not a priori, it is rendered meaningless by the weak form of verification (Copleston, 1949-50, p.73). The same outcome occurs for many statements regarding future events, descriptions of the past and vague memories; verificationsim paints acceptable statements as meaningless in its semantic attack on metaphysics.
Metaphysics is not yet vindicated. Evolving verificationist thinking saw logical positivism fade only to be replaced by Carnap’s logical empiricism. A useful distinction in metaphysical critique should be noted between theories of semantic constraint and epistemic limitation. This is a difference between what we can legitimately say about reality, and what we can know about it. Carnap (along with the positivists) charges that there is a semantic barrier beyond which we cannot coherently describe reality. In Carnap’s theory, our vocabulary plus truth determining rules create linguistic frameworks. These frameworks encapsulate meaning in rules of reference, if one seeks to speak of reality, it must be through internal questions within a given framework. If the revisionary metaphysician wishes to make ontological claims from ‘outside’ a framework, Carnap argues what they are really trying to do is ask external questions about the empirical value of adopting the specified framework (1950, pp.21-24). The enterprise of describing reality is limited to trivial questions internal to a framework e.g. ‘2+2=4?’, and instrumental external questions e.g. ‘what is the practical value of adopting geometrical vocabulary when describing objects?’ Metaphysicians claiming to be ignoring this semantic constraint would be analogised to chess players trying to make up their own rules; within the structure it’s impossible.
But is this semantic barrier not overly restrictive? It seems to place a firm limitation on the knowledge we can have of the world. It does not obviously follow that what we cannot yet understand about the world is automatically a pseudo-statement. Either this accusation is an empirical statement about the value of testability, in which case the metaphysician is open to making a counter-claim (Copleston, 1949-50, p.79), or some things must be accepted a priori. Carnap must agree to a priori reasoning in certain cases if physical theories are to be safeguarded as being the best means of describing the structure of unobservable reality. On another note, it seems the sum of all frameworks constitutes existence, yet this must be an ontological commitment in itself. The way we experience the world and thereby describe it (resulting in limitations coming from our finite conception) echoes Kant’s position on phenomenal experience. I argue than in hidden commitment to: an inaccessible noumenal world, a priori reasoning on unverifiable scientific assumptions, and the phenomenal world (cast in a semantic form as linguistic frameworks), Carnap is better thought of as a Neo-Kantian descriptive metaphysician.
Kant sought an epistemic foundation for metaphysical analysis, without recourse to purely analytic and a posteriori frameworks. Neither reason nor experience alone is sufficient to provide knowledge. The first provides content without form, the second form without content (Scruton, 1981, p.25). His synthesis produced the synthetic a priori. A form of knowledge obtained through pure reflection on ones intellectual faculties, encapsulated in recognition of how we experience the world. Analogous to a pair of spectacles one can never take off, this is the phenomenal world (Moore, 2006, p.14). Like Carnap, Kant argues the independent reality of the noumenal world is inaccessible to us, unlike Carnap, he does not see this as reason on abandon metaphysics completely. For Kant, the transcendental distinction reinforces the notion that metaphysics should exist descriptive analysis of our own concepts, as they come through perception; an ‘immanent metaphysics’. Traditional questions on such things as the divine and fundamental ontology are impossible as there is no internal potentiality to answer them. To borrow from Pirsig’s metaphor, Kant effectively discards most of the menu, in the hope that service will eventually pick up.
By refuting verificationism and reinterpreting Carnap on Kantian lines, I have come to sacrifice pure revisionary metaphysics for a descriptive analysis of our phenomenal frameworks. However, two significant problems arise from such a conception, threatening to pull the enterprise back into the realms of impossibility. Kant supported the position that reality itself is objective, yet of a character that is sculpted by the points of view through which it can be known. This epistemic limitation can be thought of as the view from possible experience (Scruton, 1981, p.20). Such a limitation serves as the foundation for the transcendental distinction between the phenomenal and noumenal worlds, meant to secure the possibility of metaphysical inquiry. But this distinction creates two fresh problems, one of unobservable structure, and one of setting the bounds of sense.
Kant reserves space for both experience and critical reflection, but over-emphasis on defining reality in relation to possible experience fails to do justice to deeper structure. The unobservable world can be utterly alien to our ordinary concepts. We need things to make sense, in the worlds of Scruton commenting on Kant, “We project our reality onto the world” (1981, p.17). For example, Euclidian geometry has been supplanted by Einstein’s theory of curved space-time. Long held assumptions over causation are being challenged by the possibility of uncaused quantum events. Special relativity is thought to imply all times are real, and we suffer from positional bias as observers. Current theoretical physics even contemplates the reality of multiple dimensions beyond our own. The reality we impose is often limited by our intellectual faculties, and can conflict with what our best science actually postulates (Ladyman, 2007, p.181). Fundamental reality obtains regardless of the phenomenal world and surreptitiously gives it structure. Surely some kind of further analysis is required to ground our concepts and safeguard the legitimacy of our metaphysics?
Furthermore, a problem of distinction must be addressed. Kant draws the boundaries of sense when separating the phenomenal from the noumenal. He argues we can have no knowledge of the noumenal world (despite the aforementioned problem of unobservable structure). But this draws the boundaries of sense from a point outside them, which if Kant is to be correct, should not even exist. To quote Wittgenstein, “in order to be able to draw a limit to thought, we should have to be able to find both sides of the limit thinkable (i.e., we should have to be able to think what cannot be thought).” (2010, Introduction). In dividing metaphysical space into two concrete realms, Kant can be criticised as succumbing to self-stultification (Moore, 2006, p.3).
So it can be said that we require an analysis of underlying structure to explain scientific explanation, and a revitalised conception of epistemic access to metaphysical analysis (rather than illegitimately drawing a Kantian distinction, or a confining semantic barrier). To answer both criticisms a final refinement must be made, metaphysics must be seen as descriptive analysis of physical structures, combined with awareness of how the implications of scientific structure interact with our phenomenal concepts. Continual advances in technology advance human capability to delve deeper into the fundamental structure of reality, moving the phenomenal world of possible experience further into the previously inaccessible noumenal world. The line between the noumenal and the phenomenal is far vaguer than Kant realised. This conception avoids setting the bounds of sense, as it portrays them as progressive.
A structure can be understood and hypothesised upon, where before was Kant’s inability to accommodate seemingly counter-intuitive scientific discoveries. Structural analysis carries the possibility of verifying metaphysical theories through their capacity to better explain physical phenomena and integrate with pre-established scientific knowledge. So a naturalised metaphysics can act as an inference to the best explanation, adding the viability of verified scientific research. Although the credence of science (especially incomplete theoretical physics) should not be haphazardly attributed to naturalised metaphysics, it can offer the benefits of ruling out untenable theories, and help describe our place in relation to underlying fundamental structure (Ladyman, 2007, p.191). As inference to the best explanation, naturalised metaphysics bears the capacity to analyse the potential ramifications of physical structures (e.g. potential randomness in Quantum Mechanics) on our own world views and interaction with reality around us. Another example would be studying the advances of neuroscience and applying the evidence to problems of free will and consciousness. As an epistemological foundation for the possibility of metaphysics, this also avoids the problem of semantic limitation. Whilst being a framework, it is only limited by our growing potentiality for epistemic access, not by a semantic system of reference in the way propagated by anti-metaphysicians. I think this barely counts as a limitation at all, let alone a problem with the power to rule out the possibility of metaphysical knowledge.
To conclude, early verifcationism failed to dismiss metaphysics, never being fully able to define its own strict criterion of meaning. Carnap failed to adequately provide a convincing semantic barrier on knowledge, although he succeeded in showing how metaphysics must be anchored in some kind of empirical framework. I compared Carnap’s and Kant’s approaches to highlight emphasis on structural limitation, favouring Kant for his broadening of the possibility of knowledge. Finally, I’ve show how an empirically informed naturalised metaphysics can fulfil our urge to reflect on our standing within fundamental realty. And that possibility of metaphysical knowledge is strongest when rooted in the epistemic foundations of progressive scientific explanation. Perhaps all Pirsig need do is realise the possibility for nourishment is there for the taking, but this restaurant operates under self-service.
Bibliography
Audi R. ‘The Cambridge Dictionary of Philosophy’ 2nd edition, (Cambridge University Press, 1999)
Ayer, A.J. ‘Language, Truth and Logic’, (Pelican Books, Middlesex, 1971)
Carnap, R. ‘Empiricism, semantics, and Ontology’, (Revue Internationale de Philosophie 4, 1950: 20-40)
Copleston, F.C. ‘The Possibility of Metaphysics’, Proceedings of the Aristotelian Society 50 (1949-50, pp. 65-82)
Hume, D. ‘An Enquiry Concerning Human Understanding’, (Oxford University Press, Oxford, 1999)
Ladyman, J. ‘Does Physics Answer Metaphysical Questions’, in Philosophy of Science: Royal Institute of Philosophy Supplement: 61 (Cambridge: Cambridge University Press, 2007, pp. 179-202)
Moore, A.W. ‘The Bounds of Sense’, Philosophical topics number 34, Nos 1 & 2, (2006)
Pirsig, R. ‘An inquiry into morals’, (New York, Bantam Books, 1991)
Scruton, R. ‘Kant’, (Oxford University Press, Oxford, 1982)
Wittgenstein, L. ‘Tractatus Logico-Philosophicus’, (CreateSpace, 2010)
