The image is from here.
Are there aspects of reality that are forever beyond the grasp of the human mind? On the optimistic view (let us call it epistemic optimism), we can eventually understand everything. In other words, our minds, aided by our sign systems, can represent any truth about reality, including the most fundamental truths of the kind traditionally sought by metaphysics. On the skeptical view (let us call it epistemic pessimism), our ability to represent reality may be irredeemably limited. As a consequence, there are truths about reality that our minds may never be able to capture.
One concept associated with this debate is that of the ineffable. To offer a provisional definition, a feature of reality is ineffable if there can be no conceptual means to represent it. The ineffable should not be confused with the unknowable. There are many things we can’t know because of various physical limitations. For instance, we can’t know exactly how many hedgehogs there were on Earth the day Barack Obama was born, or what has been going on in the regions of our universe outside our light cone. Such features are not ineffable since we have the conceptual wherewithal to represent them. It just so happens that we have no physical access to the relevant parts of reality.
What we want to know then is whether reality has any genuinely ineffable features. Progress on this issue would have immediate implications both for metaphysics and for the empirical sciences.
In a recent essay entitled “Are there ineffable aspects of reality?”, Thomas Hofweber (2016) has offered an interesting defence of what, at first glance, appears to be epistemic optimism. Hofweber makes a distinction between human-independent reality “understood as what there is” and (very Wittgensteinian-sounding) reality “understood as what is the case” (p. 166). Reality as what is the case contains all true propositions about the world. In an ingenious move based on an internalist view of our talk about propositions, Hofweber argues that the range of all propositions, including all true propositions, is a function of the human mind; more precisely, it is whatever can be stated in human language. Since we can obviously represent anything that can be stated in our language, reality as what is the case contains no ineffable facts, and reality as what there is contains no facts at all because all facts are propositions. Therefore, all facts, including any facts in the domain of metaphysics, are potentially effable.
In this essay, I want to argue that complete effability, as defended by Hofweber, is not an optimistic view of our epistemic situation. It is a potentially sceptical view that achieves a semblance of epistemic optimism by relying on a narrow definition of the ineffable. As any other version of epistemic pessimism, it is supported by the usual empirical arguments for the existence of the ineffable. These arguments are strong, and they cannot be ruled out on a priori grounds. However, I will conclude the essay by defending a moderately optimistic stance on our epistemic situation. I will attempt to show that evolutionary arguments for epistemic pessimism can also be used to undermine it.
2. Hofweber’s effability thesis
My main objection to Hofweber’s view of the ineffable does not depend on the exact details of his semantic argument for complete effability. However, I feel that it is important to sketch out Hofweber’s defence of the effability thesis in order to preempt the charge of missing the point of the argument.
The view that reality has no ineffable features can be summed up in the following proposition (Hofweber, 2016, p. 132):
(1) Everything is effable.
For several reasons, two of which I will return to later, this proposition might strike us as implausible. Surely it is presumptuous to believe that we, jittery apes with watery brains the size of a small melon, have the conceptual means to represent every single truth about reality.
However, Hofweber argues, our reluctance to believe that everything is effable is based on an assumption about the reference of “everything”. We assume “everything” to refer externally, and on this external reading “everything” ranges over all elements of reality outside our heads. According to Hofweber, this assumption might be wrong. It may be the case, he argues, that “everything” refers internally. In other words, instead of ranging over all elements of reality, it ranges over all possible linguistic propositions of the form “[It is the case] that p”. But if “everything” refers internally, then (1) is not only plausible; it is necessarily true. Every possible proposition of the form “[It is the case] that p” is indeed effable for the simple reason that it will be, or has been, or could be, conceptually represented by means of our language. Therefore, everything is effable, and reality as what is the case has no ineffable features. Quod erat demonstrandum.
As I have already indicated, it is not my intention to question Hofweber’s semantic argument for the effability thesis. In fact, I am happy to grant that his conclusion is correct: given his definition of the ineffable and the internalist reading of “everything”, the effability thesis is true, and everything is thus effable.
I will, however, argue that Hofweber sidesteps the problem by defining the ineffable out of existence. As I see it, he achieves this result in at least two ways: by limiting conceptual content to linguistic content and by excluding the incomprehensible from the realm of the ineffable. Dissolving a problem in this way is often a helpful philosophical strategy, but I doubt that, in the case of the ineffable, it is justified. Hofweber himself admits that his solution may feel like an “unsatisfactory”, “shallow victory for the effability thesis” (p. 140), though he clearly believes that this sense of dissatisfaction is illusory. In the next section, I will try to show that this dissatisfaction is based on legitimate concerns.
3. Defining the ineffable
In the introduction, I provisionally defined the ineffable as any features of reality that can’t be represented for lack of conceptual means. At least three elements of this definition require further comment. One of them is a missing element: so far I have used the passive voice (“can’t be represented”) and an existential clause (“there is no conceptual means”) to avoid mentioning who it is that can’t eff certain features of reality. This omission stands in stark contrast to Hofweber’s explicit focus on what is ineffable for “us”, “humans” and “human minds”. I will revisit this missing element of the definition in the last section.
The other two elements that need closer attention are conceptual means and feature of reality. I will now look at each of them in turn.
What does it mean to represent something conceptually? To answer this, I take my cue from Hofweber’s remark on mystical insights: these, he notes, are claimed to be not mere “feelings” but rather “thoughts with contents that can be true or false” (p. 128). Quite regardless of whether mystical insights do indeed contain such thoughts, I believe that this is a good point of departure. Whatever else might be said about conceptual representations, they must fulfil at least three basic stipulations:
(a) They are thoughts: they are goings-on in a mind.
(b) They are about something: they have content.
(c) They can be true or false: their content can either match or mismatch the rest of reality.
Can this understanding of conceptual representations as thoughts that can have true or false content help us clarify the notion of the ineffable? I believe that it can. Before moving on, however, let us take a closer look at stipulation (a). Perhaps it smuggles in a ticking phenomenal bomb that can render the whole edifice empirically useless. Perhaps the notion of goings-on in a mind comes perilously close to that of a phenomenal concept.
We can plausibly grant that phenomenal concepts can be true or false: if I hallucinate a big fluffy tail growing out of the end of my spine, it seems fair to say that I have a false representation of my body, constituted by one or more phenomenal concepts. However, the falsity of my tail hallucination can be observed by others and mediated by non-phenomenal means. Someone can say: “It is not the case that Konstantin has a big fluffy tail”; or they can take a picture of my tailless behind.
The trouble begins if we allow for genuinely private phenomenal concepts that can’t be shared with anyone except their experiencer. Suppose that I suddenly experience a phenomenal representation of a highly mysterious feature of reality that I can neither describe in words nor capture with my smartphone or detect with an array of laser interferometers. I may be fully convinced of the profound truth of my experience, or I may believe it to be a weird phenomenal fluctuation that has nothing to do with the part of reality that is not me. Either way, I need to find a shareable conceptual representation of my experience if I want to have a meaningful exchange concerning its truth or falsity with anyone else. Like private language, genuinely private phenomenal concepts seem vaguely conceivable, but it is not clear what we could ever do with them philosophically, let alone scientifically. This is why a fourth stipulation seems necessary:
(d) Conceptual representations (i.e. mental goings-on that have content) are shareable with other minds.
This echoes Hofweber’s requirement that the ineffable be suitably coarse-grained – in other words, that something be considered ineffable only if “it can’t be represented no matter who, where, and when you are” (p. 129). However, Hofweber’s reasons for positing this are at least partly different from mine, and I will try to show that this difference is important.
Hofweber insists on a coarse-grained view of the ineffable in order to rule out trivial ineffable truths. To modify Hofweber’s example of such trivial ineffability, if the fact that I feel exhilarated by the first snowfall is distinct from the fact that Konstantin feels exhilarated by the first snowfall, then the true nature of this distinction is ineffable for anyone except me.
I agree that I have unique phenomenal access to my sense of exhilaration. At the same time, I don’t think that this difference in access makes my first-person perspective ineffable for every other human being, either in a trivial sense or in any other sense. How exactly I (as opposed to Konstantin) feel when it starts snowing in November is unknowable for every human being except me in more or less the same sense as it is unknowable for every single human being how many hedgehogs inhabited our planet at the moment of Obama’s birth. By the same token, just like we can, as a civilisation, roughly estimate the size of the hedgehog population on 4 August 1961 based on the evidence we have today, I can rely on the goings-on in other people’s minds to represent publicly the difference between a third-person and first-person view of my snow-induced exhilaration. This is why I could discuss it in a philosophy paper, or I could paint it on a piece of canvas, or I could make an arthouse film about the tragedy of being an isolated Dasein in a snowy world of isolated Daseins.
It is likely that none of these representations would cause anyone to have exactly the same goings-on in their brain as I have in mine. Indeed, some of the resulting goings-on in other minds might be wildly different from mine. However – and this is a crucial point – it is also likely that no mental going-on is ever qualitatively, let alone numerically, identical to another. The conceptual representation in the receiver’s mind will not be the same as in the transmitter’s mind because conceptual representations are not simply moved from mind to mind; they are recreated from scratch at the receiving end every single time. Any structural similarity between two representations will ultimately be a factor of the structural similarities between the transmitter and the receiver, as well as between their respective environments and histories. But if that is the case, it doesn’t matter what is being transmitted: the tragic phenomenal solitude of the self, the average size of the human pancreas, or the Pythagorean Theorem. What is perhaps even more important in the context of this essay, it doesn’t matter how it is being transmitted: by means of linguistic propositions, through algebraic notation, as a string quartet in C minor, or in some other way that is not (yet) available to humans. The key question is whether the means of conceptual sharing employed by the transmitter can trigger structurally similar goings-on in the mind of the receiver.
In “Are there ineffable aspects of reality?”, Hofweber repeatedly indicates that the ineffability question concerns both what we can “say” and what we can “think”, or have a “conceptual representation” of. His discussion of propositions and the linguistic focus of his main argument, described in the previous section, suggest that he expects everything that is the case to be linguistically mediated. Even if non-linguistic thought is possible, and even if non-linguistic thought can have some kind of content, this content must be translatable into a linguistic proposition (specifically, a that-clause) if it is to say anything at all about what is the case. I assume that it is for this reason that Hofweber describes the unique aspects of a first-person view as ineffable, even if trivially so. After all, most of us feel that non-linguistic experience can’t be given the form of a linguistic proposition without losing some crucial aspect of its experiential-ness.
If my discussion above succeeded in triggering the right kind of goings-on in the reader’s mind, it must be becoming clear by now that I believe the process of reality-effing to go the other way. To ask whether we have conceptual means to represent a feature of reality is not to ask if we can translate our insights into propositions. Instead, to eff a feature of reality is to use whatever means possible, including language, to engineer an insight, i.e. a thought with content that is true and can be shared.
On this view, it is still possible to observe, along with Hofweber, that what is the case is “not part of reality [as what there is] in the sense that there is no domain of propositions or facts of which we might capture more or fewer” (p. 165). It seems likely that any shareable content that we have in our heads, including content that can be true or false, emerges in our heads, and reality as what is the case is thus range-dependent on our representational capacity. Having said that, it is also trivially true that our conceptual representations are a part of reality as what there is. This is because we, including whatever we have in our heads, are a part of reality as what there is. For this reason, focusing on language in general and on the internalist or externalist picture of the propositional in particular might not be particularly helpful if our goal is to decide between epistemic optimism and epistemic pessimism. The important question is not whether truths about reality are in our heads or “simply there” (p. 161). It is the rest of reality – the Non-Us part of reality – that is simply there. Given that our species is still around, there must be at least some matches between the Us-part and the Non-Us part, i.e. there must be some productive structural connection between what goes on in our heads and what goes on in the rest of reality as what there is. The question is then how far these matches can possibly reach. Are there features of reality to which our brains, however thoroughly racked, simply can’t form a structural link?
All this might sound very metaphorical and vague in a suspiciously continental kind of way, especially compared to a nicely formalised discussion of quantifier reference. However, my hope is that, if we approach the problem of ineffability as I did above, then focusing on language-mediated content to the exclusion of other possibilities will feel a bit more like looking for your lost keys under the proverbial streetlight because that is the only spot where you can actually see.
As soon as we allow that the world as die Gesamtheit der Tatsachen (Tractatus, 1.1) might be representable by means other than human language, evolutionary arguments for the existence of ineffable features of reality are no longer sidestepped, and they come back with a vengeance. Arguing that all representable truths are those that can be represented by human language once again begins to feel like claiming the same for any animal communication system. After all, if every truth we can ever discover will be a linguistic proposition, in what sense is our current epistemic situation fundamentally different from that of the honeybee?
To be sure, the combinatorial range of currently existing human languages is immensely richer than the combinatorial range of the waggle dance. Clearly, we can convey information about the spatial location of a source of nectar, while bees can’t waggle-dance Plato’s ruminations on the nature of piety or even how a cat is different from a panda. We could of course manipulate a bee’s dance in some way to expand its representational range, perhaps by hooking her up to some kind of remote control; but that way we can only succeed in using the bee as an uncomprehending tool to represent Plato and cats to ourselves, not to the bee or her sisters. We simply can’t induce certain goings-on in a bee brain unless we radically upgrade its hardware and contents.
But if this picture is true of bees, there is hardly any a priori way to show that it can’t be true of us. We might never be able to understand at least some of the truths effed by sufficiently advanced alien intelligence. The aliens could of course remote-control us into “representing” their ineffable truths, but just like bees waggle-dancing Euthyphro, we would be none the wiser. Indeed, the aliens could, as Hofweber imagines, give us some names for ineffable features of reality, presumably adjusted to our orthography or phonology and complete with lengthy explanations relying on human analogies and metaphors. Even with all that help, we might be unable to have the required kind of mental goings-on simply because, due to our evolutionary history, we might lack access to whatever mental content allows the aliens to “get” it. Instead, we might end up acting out a version of Searle’s Chinese Room, passing around strings of symbols that we think we understand but without ever getting at the alien concepts behind them. The aliens’ only hope of making us “get” it would then be to upgrade our cognitive hardware and its contents.
As far as I can see, the only way to rule out such a scenario by philosophical means is to make “ineffable” strictly distinct from “incomprehensible”, which is exactly what Hofweber does (p. 131). This move also allows Hofweber to argue that ineffable features of reality, if any, appear to be suspiciously well hidden from us, so that “for all intents and purposes every fact we encounter is effable”, and the ineffable is therefore apparently “irrelevant for our attempts to understand various parts of the world in inquiry” (p. 144).
This claim presents a good opportunity to clarify another element of my definition of the ineffable. What is a feature of reality? If it is whatever can be given a name, then Hofweber is right: we have never encountered anything ineffable. Suppose that tomorrow an alien spacecraft lands in front of the Carolina Rediviva library in Uppsala. A team of alien scientists emerge from behind a sliding door and announce in flawless Swedish that they have long since cracked all the mysteries that keep human philosophers exercised and at times even gainfully employed. Suppose also that one of the bespectacled earthlings who were passing by hasn’t been entirely bowled over by awe and euphoria and manages to ask some questions. “How splendid!” she exclaims. “What is the correct interpretation of quantum mechanics, and how come conscious experience can emerge from unconscious stuff?” “Oh, that’s easy,” the aliens reply. “The answer to your first question is den grömtliga bråbbelknycken, while the answer to your second question is ett föremustsamt bunderspräpp.” It is quite likely that the bespectacled earthling will be left at least as unsatisfied by this reply as one might be by Hofweber’s definition of the ineffable. Clearly, the aliens could simply have said “42” and been done with it. Indeed, they could have played the earthling’s questions back at her and then declared, “Why, it’s completely effable even with your measly intelligence! Why don’t you ask us about something that you really can’t represent conceptually?”
Now, I do not wish to claim that the correct metaphysical interpretation of quantum mechanics or a solution to the hard problem of consciousness are beyond our current representational capacity. But I have been convinced by people like Colin McGinn that at the very least this is a distinct possibility. This is why I find it especially interesting that Hofweber mentions McGinn as an example of someone who is “concerned with the ineffable in these other senses” (p. 156, footnote); that is, in the senses that cover features of reality we can point at without knowing how to represent them, or if we have the conceptual resources to represent them in the first place. Whatever feature makes the phenomena described by quantum mechanics compatible with the classical world of our everyday experience is one example; whatever feature combines the conscious and the unconscious in one universe is another. In other words, if we collapse the ineffable/incomprehensible distinction, as I believe we should, it is not difficult to find cases where we might already have run up against a locked epistemic door, behind which there may well be something ineffable. In fact, allowing for the possibility of ineffable features that we can’t even point at, every single act of human perception may be an act of encountering an epistemic door of this kind – just as it surely is for bees. What is more, there is nothing particularly esoteric about this idea: that there is more to reality than ever meets the eye is a time-honoured staple of human intellectual history.
To sum up my arguments so far, I believe that we can do justice to the ineffability question by defining the ineffable as follows: the ineffable is whatever part of reality as what there is that cannot be matched up with reality as what is the case by means of any shareable mental goings-on that have content and can be true or false. On this definition, evolutionary arguments for the existence of the ineffable for the biological species homo sapiens sapiens seem fairly convincing. We have no a priori reason to believe that our ability to eff reality is so fundamentally different from that of other species that we can expect to eff the whole of reality.
Admittedly, this is a form of epistemic pessimism. However, as I will argue in the last section of the essay, it is no more pessimistic than Hofweber’s effability thesis. Indeed, it might even give us some cause for hope.
4. Moderate epistemic pessimism
Early in his essay, Hofweber presents some common arguments in support of the possibility of the ineffable. Needless to say, these arguments are not conclusive – few arguments are – but I believe they are sufficiently strong. In particular, the argument from built-in cognitive limitations and the argument from analogy with other biological species make epistemic pessimism the more plausible option, as long as we don’t preemptively identify what is the case with what can be stated by humans in human language.
I don’t know if Hofweber views his effability thesis as a kind of epistemic pessimism. Given his remarks at the Ontology and Metaontology workshop (Uppsala University, 15 November 2017), it could go either way. “What exists is one thing; what is true is another thing” can be read as an expression of profound pessimism, while “we are not limited [in our representational capacities] if we understand reality as the totality of facts” probably allows an optimistic interpretation. My own view is that Hofweber’s take on effability is pessimistic as long as it presupposes that our cognitive make-up will remain more or less the same. This is not to deny that human language can be modified and used by beings more advanced than us to mediate conceptual representations of what is ineffable for modern humans. As I tried to show in the previous section, aliens with superior reality-effing abilities might be able to employ any given human language to mediate goings-on in their minds, just like we could modify the waggle dance of a bee and use it (perhaps by controlling an actual bee, or by using a computer-generated bee image) to represent to ourselves anything from Plato’s dialogues to the role of the Higgs boson in the Standard model. However, as I have also argued, even if the aliens taught us this modified version of our language, our brains might never generate anything sufficiently similar to their mental goings-on due to some structural limitation.
Having said all that, I would like to end this essay on an upbeat note. I believe that we do not have to assume that our cognitive make-up will forever remain the same. To be sure, the fact that we are just one animal species among many, shaped by a process whose driving force has been the struggle for survival rather than a desire to uncover the foundations of reality, strongly suggests that our brains have cognitive limitations. However, the same fact also suggests that our evolution is not over. Why would it be? If anything, it is probably speeding up, both in terms of traditional selection pressures and as regards our enormously complex extended phenotype, also known as culture. Perhaps even more importantly, our technological development might result in significant cognitive enhancements of human organisms, or indeed in creation of new sentient organisms, whether biological or non-biological. I don’t quite share all the rosy visions of some transhumanist thinkers, but leaving aside any ethical implications of such developments, it is at least possible that our descendants will be able to wrap their structurally different brains around something currently ineffable. Consequently, they might be able to understand the features of reality underlying our most intractable metaphysical puzzles, just as we understand the features of reality underlying rain and thunder.
This finally brings me to the missing element of my definition. As I was trying to define the ineffable in the previous sections, I hesitated to talk about “us” or even about “humans”. A hundred thousand or a million years from now (always assuming that our civilization does not self-destruct), will our descendants – especially non-biological ones – be “us” or “human”? And do we, for that matter, count as “us” our distant ancestors from several million years ago? After all, it is quite likely that they wouldn’t be able to eff much of what we now understand even if we built a time machine and went back to try and explain it to them.
The best answer I can give at this point is that it is a matter of how we choose to interpret “us”. If we choose to identify as “us” our evolutionary ancestors and descendants, it is at least conceivable that our reality-effing capacity will dramatically increase with time. If we do not, then our ability to match the Non-Us part of reality as what there is might be fixed. Either way, this is an empirical question. All we can ever do is use whatever means we have to trigger insights in our watery or post-watery brains.
 It is perhaps worth noting that a conceptual representation, i. e. the content of a thought, can miss the rest of reality either partly or completely. If it misses partly, it misrepresents a feature of reality. If it misses completely, it represents no feature of reality except itself.
 At the Ontology and Metaontology workshop (Uppsala University, 15 November 2017), Hofweber remarked that the effability thesis “does not apply to non-conceptual representation”, which again seems to imply that he views conceptual representation as mediated by, or at least translatable to, human language.
 I’m not sure about committed eliminativists.
 Here I ought to clarify that I don’t buy the distinction between propositions and their linguistic form. Instead, I have always thought of it as a sneaky trick that everyone uses to link that-clauses to the joints of external reality, or at any rate make propositions appear more important than they are. In any case, even if Hofweber does assume this distinction to be important, I would argue that he completely ignores it in his main argument. In other words, whenever Hofweber talks of propositions, I see no evidence that he has in mind non-linguistic propositions.
 Just to be clear: the focus on mental goings-on, whether shareable or not, is mine rather than Hofweber’s. Part of my intent is to contrast it with Hofweber’s view of thoughts as propositions.
 I really wish I had asked him at the workshop.
Hofweber, T. (2016). “Are there ineffable aspects of reality?” Oxford Studies in Metaphysics 10: 124–170.