Experiencing humility: hope for public debate?

When I was young (yes, stop snickering), when I was young I was often amazed at people’s knowledge. Most people had opinions about everything. The government issued a statement about a new policy and my father or one of my uncles already knew that the policy wouldn’t work. This admiration didn’t stop during my adolescence: I remember listening in awe when friends saw through all the motives and consequences of political decisions. How did they figure it all out? – Well, they probably didn’t. Or not much of it. I don’t want to sound condescending but most of us probably don’t understand the implications of political decisions all that well. Yet, judging by the readiness and vehemence of our contributions to public debate, most of us do at least give the impression of relative expertise. If this is correct, there is a disproportion between actual understanding and confidence in our opinions. In what follows, I’d like to suggest that amending this disproportion might hold the key to improving public debate.

According to Kees van den Bos and other social scientists,* this disproportion is one of the crucial factors leading to polarisation in public debate. However, the inverse also seems to be true: if people are asked to explain how certain policies work and experience that they understand these policies less well than they thought, they are likely to exhibit more moderation in their views of these policies. Fernbach et al. (2013) write:

“Across three studies, we found that people have unjustified confidence in their understanding of policies. Attempting to generate a mechanistic explanation under-mines this illusion of understanding and leads people to endorse more moderate positions. Mechanistic explanation generation also influences political behavior, making people less likely to donate to relevant advocacy groups. These moderation effects on judgment and decision making do not occur when people are asked to enumerate reasons for their position. We propose that generating mechanistic explanations leads people to endorse more moderate positions by forcing them to confront their ignorance. In contrast, reasons can draw on values, hearsay, and general principles that do not require much knowledge.”

So while I might become increasingly stubborn if you ask me to give reasons for p, I might become more moderate if you ask me to explain how p works. According to the researchers, this is the case because in the latter scenario I am humbled by experiencing the limits of my knowledge. I guess it won’t be too much to ask you to imagine examples. Asking how certain policies of, say, traffic regulation or migration work in practice might even lead politicians themselves to moderation.

What precisely is it that leads to moderation? My hunch is that the effect is produced by experiencing humility. This means that it is vital that the subject in question experiences their lack of knowledge. It is probably no good if I am told that I lack knowledge. (In fact, I believe that this might instil resentment.) The point is that I realise my lack in my own attempt at an explanation. So what I would like to emphasise is that the moderating effect is probably owing to experiencing this lack rather than merely knowing about this lack. Of course, I know that I don’t know how precisely certain policies work. But it’s still quite another thing to experience this ignorance in attempting to explain such policies. In other words, the Socratic attitude alone doesn’t help.

If this effect persists, this finding might indeed help ameliorating conversations and debates. Instead of telling people that they are wrong or asking for reasons, we might simply ask how the proposed idea works. This requires of course humility on part of all interlocutors. A good start might be debates in philosophy.


* I am grateful to Hendrik Siebe, Diego Castro and Leopold Hess for conversations about this work online and offline.

Diversifying scholarship. Or how the paper model kills history

Once upon a time a BA student handed in a proposal for a paper on Hume’s account of substance. The student proposed to show that Hume’s account was wrong, and that Aristotle’s account was superior to Hume’s. If memory serves, I talked the student out of this idea and suggested that he build his paper around an analysis of a brief passage in Hume’s Treatise. – The proposal was problematic for several reasons. But what I want to write about is not the student or his proposal. Rather I want to zoom in on our way of approaching historical texts (in philosophy). The anecdote about the proposal can help to show what the problem is. As I see it, the standard journal article has severe repercussions on the way we teach and practise scholarship in the history of philosophy. It narrows our way of reading texts and counters attempts at diversification of the canon. If we want to overcome these repercussions, it will help to reinstate other forms of writing, especially the form of the commentary.

So what’s wrong with journal articles? Let me begin by saying that there is nothing wrong with articles themselves. The problem is that articles are the decisive and almost only form of disseminating scholarship. The typical structure of a paper is governed by two elements: the claim, and arguments for that claim. So a historian typically articulates a claim about a text (or more often about claims in the secondary literature about a text) and provides arguments for embracing that claim. This way we produce a lot of fine scholarship and discussion. But if we make it the leading format, a number of things fall through the cracks.

An immediate consequence is that that the historical text has the status of evidence for the claim. So the focus is not on the historical material but the claim of the historian. If we teach students to write papers of this sort, we teach them to focus on their claims rather than on the material. You can see this in the student’s approach to Hume: the point was to evaluate Hume’s account. Rather than figuring out what was going on in Hume’s text and what it might be responding to, the focus is on making a claim about what is the supposed doctrine. The latter approach immediately abstracts away from the text and thus from the material of discussion. What’s wrong with that? Of course, such an abstract approach is fine if you’re already immersed in an on-going discussion or perhaps even a tradition of discussions about the text. In that case you’re mainly engaging with the secondary literature. But this abstract approach does not work for beginners. Why? Arguably, the text itself sets constraints that have to observed if the discussion is to make sense. What are these constraints? I’m not saying they are fixed once and for all. Quite the contrary! But they have to be established in relation to the text. So before you can say anything about substance in Hume, you have to see where and how the term is used and whether it makes sense to evaluate it in relation to Aristotle. (My hunch is that, in Treatise, Hume rejects the Aristotelian idea of substance altogether; thus saying that Aristotle’s notion is superior is like saying that apples are superior to bananas). The upshot is: before you can digest the secondary literature, you have to understand how the textual constraints are established that guide the discussions in the secondary literature.

What we might forget, then, if we teach on the basis of secondary literature, is how these constraints were established in the long tradition of textual scholarship. When we open an edition of the Critique of Pure Reason, we see the text through the lens of thick layers of scholarship. When we say that certain passages are “dark”, “difficult” or “important”, we don’t just speak our mind. Rather we echo many generations of diligent scholarship. We might hear that a certain passage is tricky before we even open the book. But rather than having students parrot that Kant writes “difficult prose”, we should teach them to find their way through that prose. That requires engagement with the text: line by line, word by word, translation by mistranslation. Let’s call this mode of reading linear reading as opposed to abstract reading. It is one thing to say what “synthetic apperception” is. It’s quite another thing to figure out how Kant moves from one sentence to the next. The close and often despair-inducing attention to the details of the text are necessary for establishing an interpretation. Of course, it is fine to resort to guidance, but we have to see the often tenuous connection between the text and the interpretation, let a lone the claim about a text. In other words, we have to see how abstract reading emerges from linear reading.

My point is not that we shouldn’t read (or teach what’s in the) secondary literature. My point is that secondary literature or abstract reading is based on a linear engagement with the text that is obscured by the paper model. The paper model suggests that you read a bit and then make a fairly abstract claim (about the text or, more often, about an interpretation of the text). But the paper model obscures hundreds of years or at least decennia of linear reading. What students have to learn (and what perhaps even we, as teachers, need to remind ourselves of) is how one sentence leads to the next. Only then does the abstract reading presented in the secondary literature become visible for what it is: as an outcome of a particular linear reading.

But how can we teach linear reading? My suggestion is quite simple: Rather than essay writing, students in the history of philosophy should begin by learning to write commentaries to texts. As I argued earlier, there is a fair amount of philosophical genres beyond the paper model. At least part of our education should consist in being confronted with a piece of text (no more than half a page) and learning to comment on that piece, perhaps translating it first, going through it line by line, pointing out claims as well as obscurities and raising questions that point to desirable explanations. This way, students will learn to approach the texts independently. While it might be easy to parrot that “Hegel is difficult to read”, it takes courage to say that a concrete piece of text is difficult to understand. In the latter case, the remark is not a judgment but the starting point of an analysis that might allow for a first tentative explanation (e.g. of why the difficulty arises).

Ultimately, my hope is that this approach, i.e. the linear commentary to concrete pieces of text, will lead (back) to a diversification of scholarship. Of course, it’s nice to read, for instance, the next paper on Hume claiming that he is an idealist or whatever. But it would help if that scholarship would (again) be complemented by commentaries to the texts. Nota bene: such scholarship is available even today. But we don’t teach it very much.

Apart from learning how to read linearly and closely, such training is the precondition of what is often called the diversification of the canon. If we really want to expand the boundaries of the canon, the paper model will restrain us (too much) in what we find acceptable. Before we even open a page of Kant, our lens is shaped through layers of linear reading. But when we open the books of authors that are still fairly new to us, we have hardly any traditions of reading to fall back on. If we start writing the typical papers in advance of establishing constraints through careful linear reading, we are prone to just carry over the claims and habits familiar from familiar scholarship. I’m not saying that this is bound to happen, but diligent textual commentaries would provide a firmer grasp of the texts on their own terms. In this sense, diversification of the canon requires diversification of scholarship.

An ethics of climate change?

Imagine that you watch someone putting down a substance on all the playgrounds in your neighbourhood. You are then reliably informed that the substance is poisonous and that it has been put down on all playgrounds.

What would you do? Would you keep quiet about it? – Perhaps you should. Even if you see some children die, you don’t know for sure whether it’s really the poison that is lethal, do you?*



* Some comments:

— Faced with the question of anthropogenic climate change, most far-right parties I know endorse a variant of this argument. Are there any other options? Probably: (a) Some mainstream parties in Germany worry most about the issue that the guy putting down the poison might be out of a job if we cause a fuss. (b) Other parties campaign for building new playgrounds. (c) A further option is to ask the producers of poison for the best strategy. (d) Oh, and another commonly favoured move is to dismiss protests of children as ignorant. Did I miss something?

— If you’ve read this far, why not sign this petition by Martin Kusch (Vienna): “Philosophers for Future”?

— While I think that dismissing findings of climate sciences or other concerns about climate change is immoral, I also see that ‘spreading the word‘ is rather difficult. Meehan Crist writes: “On a rapidly warming planet, one function of climate writing is to get the word out—to spark and help shape public discourse in the midst of ongoing and accelerating catastrophe. We are already too late to prevent some degree of unprecedented change. We know it’s going to be bad, but human activity today could still make the future worse. So it’s true both that we are too late and that there is no time to be lost. Yet if we get the framing of this story wrong—if we see the issue as a matter of individual consumer choice, for example, or choose a purely emotional rather than an explicitly political framing—we risk missing the point altogether.”

Social syllogizing. William of Ockham, John Lutterell, Walter Chatton and Adam Wodeham on the division of epistemic labour*

In my last post, I suggested that Ockham’s account of language and concepts might be founded on the thesis of a “division of linguistic labour”. In what follows, I would like to say more about this thesis by showing how it figures in the debate over the demonstrability of articles of faith. Far from being a debate of merely theological interest, I believe that the thesis is put to work as founding the idea of a division of epistemic labour, even between the living wayfarers (viatores) and those considered happy in heaven (beati). [If you are mainly interested in Ockham’s account of the division of epistemic labour, you can jump straight to section 2 of this post.]

When Ockham’s pupil Adam Wodeham delivered his Commentary on Peter Lombard’s Sentences around 1330 in Norwich, Ockham had already been excommunicated. But this didn’t stop Wodeham from defending and elaborating Ockham’s doctrines. Like many of his contemporaries, he particularly dealt with questions concerning the relation between theology and the Aristotelian notion of science. One of these questions was whether articles of faith such as ‘God is three and one’ can be scientifically demonstrated. Wodeham’s answer to this question might sound quite peculiar: Instead of a clear ‘yes’ or ‘no’, he replies that it is possible in principle to demonstrate such creditive propositions, but not by means that are available to wayfarers, i.e. human beings in this life. Yet they are demonstrable through sentences formed by one who is blessed in heaven. So, in order for this solution to work, we have to imagine a syllogism that contains the premises of a blessed one in heaven and the conclusion of a wayfarer in this earthly life. In this sense, the epistemic labour is shared even across domains as different as heaven and earth …

Wodeham provides intriguing arguments to defend this thesis, which was originally proposed by Ockham. But what is at stake here? To be sure, it’s one of the big themes: the relation between faith and reason. The question of the demonstrability of creditive propositions aims at the epistemological status of theology. Quoting Augustine, already Peter Lombard warned theologians against endorsing the truths of faith on the grounds of ecclesiastical authorities alone. If theologians do not wish to be utterly defenceless against hair-splitting logicians, they had better come up with some reasonable arguments as well. But how was the gap between natural reason and faith to be bridged? From the mid 13th century onwards, Aristotle’s Posterior Analytics – a book that clearly did not consider matters of catholic faith – was the leading source on scientific knowledge. If the traditional talk of ‘truths of faith’ is to be taken seriously and that is taken literally, then we require a general answer to the question of what it means for such a ‘truth’ to be true and demonstrable. Yet, if articles of faith were as evident as principles of science, then faith would have to belong to the realm of naturally accessible human knowledge. And since the knowledge of both realms would be equally evident, the difference between the wayfarer’s faith and the beatific vision of a blessed one would be lost. If, however, articles of faith belong to the realm of heavenly knowledge only, then it is incongruous to call theology a science.

I think that Ockham’s solution to this dilemma gave a new meaning to the debate, in the sense that it is rooted in the conviction that knowledge – no matter whether it’s heavenly or earthly – has a general form: it is subject to general criteria of rationality. This form or structure is for Ockham psychologically real in that knowledge is generally given in the structure of a mental language, which means (to a first approximation) that the concepts we have in our minds are structured in ways significantly similar to the sentences we utter. Thus, the difference between a wayfarer (viator) and a blessed one (beatus) is not that they have different types of lower and higher knowledge (or rationality), the difference is simply that they are in different epistemic situations. But for Ockham’s solution to work, it is not enough that there is a common structure of knowledge. We also have to take the different epistemic situations of different knowers into account. Thus, the idea of a common rationality requires a division of epistemic labour. Let’s now look at the debate.

  1. Lutterell’s critique

Ockham’s teaching in Oxford and London provoked many criticisms. Already John Lutterell (Chancellor of Oxford University till 1322) was deeply concerned about Ockham’s doctrines: Ockham held that we do not know things but mental sentences whose terms stand for things. Mental terms, in turn, are caused by cognizing things. This doctrine tries to acknowledge the fact that an assent or dissent does not aim at things but at propositions: we don’t say, for instance, ‘the stone is true’, but ‘the mental sentence’, whose subject-term stands for a stone, ‘is true’. To have knowledge, then, is to have a certain mental attitude towards a proposition. Lutterell, however, concluded: “This is dangerous, because it eliminates all real science and faith.” (Libellus contra Ockham, art. 30)

Now since, on Ockham’s account, forming an evident judgement requires epistemic access to the things the proposition refers to, Ockham indeed denied that theology is, strictly speaking, a science and he denied that articles of faith can be demonstrated in this life, although they could be demonstrated in principle by means of premises formed by someone who has epistemic access to God, namely a blessed one. Lutterell, however, won’t have that. One of his main arguments against Ockham’s position is that the blessed would not form a proposition at all. On his account – and in accordance with many medieval theologians resorting to Augustine – the blessed one has a sort of vision which enables him or her to form a single and simple unstructured concept (Libellus contra Ockham, art. 4), whereas humans in this life attain simplicity only in complex ways, that is they do not attain real simplicity at all.

An helpful illustration of the difference between complex and simple thought is the difference between writing a text on your word-processor and looking at the photocopy or pdf of a text: With the word-processor you can alter the text you’re writing, you can copy words or phrases and put them elsewhere and so on; by looking at the photocopy you get the whole text at once, but there’s no way of altering the text or transporting bits of it elsewhere.

Simplicity is a feature of divine perfection and is therefore traditionally held to be a property of the beatific cognition: since, unlike complexity which is often associated with materiality, spatial and temporal order, simplicity precludes falsity or error. Seeing God, then, precludes the formation of complex concepts or any kind of discursive thinking. On Lutterell’s account, the knowledge of the blessed is shaped with regard to the object of knowledge and is thus simple, whereas the wayfarer has to form complex propositions about God, hence reaching the simple only through complex means.

Lutterell’s defence of theology as a science is largely owing to Aquinas, who held that we can justly claim that our complex conclusions about the divine are true, so long as we believe in the principles of the higher and simple knowledge of the blessed ones. But Ockham didn’t buy this argument, maintaining that we cannot know conclusions unless we evidently know the premises.

  1. Ockham’s position

Ockham’s own position is quite different. On my reading (see here and here for details), his point is that this way of talking and dividing knowledge is generally ill-formed. So his arguments amount to saying: I grant that these articles of faith aren’t demonstrable in this life, since otherwise they wouldn’t be articles of faith, and there would be no difference between this life and beatific vision. I also grant that we have to construe articles of faith as being demonstrable in principle, since otherwise our faith would aim at nothing. But if we want to say that our sentences could be demonstrable through the premises of the blessed, we first have to grant that our and their knowledge is compatible. And since our knowledge is structured, we had better assume that their knowledge is structured as well, otherwise it wouldn’t be compatible with ours.

But how can Ockham reasonably hold that articles of faith are demonstrable in principle? A scientifically demonstrable conclusion requires that the proposition is (1) necessary, that it is (2) deducible from necessary and evident propositions in a syllogistic discourse and that it is (3) doubtful (otherwise it would be self-evident, and therefore not demonstrable from other propositions). Now, the blessed one can form the required evident propositions, but the trouble is, the blessed one couldn’t form a doubtful proposition, since it would be – presumably – immediately evident to him or her, and thus independent of any argument or any other proposition.

To fulfill the criteria of demonstration it is required that we are in a position to form a doubtful proposition which the blessed one could access in principle. Ockham argues that this requirement is met, if there is someone who can doubt the proposition or if the blessed one was once in a position to doubt the proposition, which in turn requires that the sentence structures are preserved.

As I see it, Ockham’s reply combines two lines of argument: On the one hand, he insists that the structure of thought is the same for all. On the other hand, he argues that the dubitability is provided once there is someone (else) in a position to doubt (see In I Sent, d. 3, q. 4; OTh II, 440-441). It is with this move that Ockham makes the case for a division of epistemic labour. Properties of thought do not have to be fulfilled by every individual thinker. Rather it suffices if someone can fulfil them. In the given case, demonstrability even requires that thinkers are (or have) been in different epistemic situations.

  1. Chatton’s critique

Ockham’s idea is, then, that the very same proposition is doubted or believed by the wayfarer and evidently known by the blessed one. So it is possible that someone, let’s call her Martha, reads a theological treatise and stumbles upon the sentence ‘God is three and one’. Martha thinks hard, but neither can she infer this sentence from any principles she knows, nor did she have any experience with any entity that is at once three and one. We may say that she simply entertains the sentence in her mind, without judging its truth or falsity. But then she thinks even harder and realises the overt contradiction in the ascription of something being three and one at once. Now Martha, not familiar with the many logical twists and turns of trinitarian debates, concludes that it may be impossible for something to be three and one at once and doubts the sentence. Now, she may have a conversation with the author of the theological treatise in question and find him rather learned and trustworthy, so that she eventually casts aside her doubts and believes that after all there might be one entity that is very special just because it is beyond contradictions in terms. We might say then that she believes the sentence on certain premises. Now, let’s assume that Martha dies and is entitled to beatific vision. She might then, as it were, remember the sentence and evidently judge that it is necessarily true.

Although one might argue that this judgement is not a demonstration, because the kind of sentences she can form on the grounds of immediate experience are self-evident and not dubitable, we can reply with Ockham that it is indeed a demonstration because the blessed one’s newly acquired sentences are the very premises from which she infers the truth of the very sentence she formerly believed. Sentences and its parts are transportable. They can recur in different contexts, even in heaven. But Ockham pushes the point even further: it is sufficient that someone can doubt the proposition in question. Presumably, the blessed ones cannot doubt articles of faith. Therefore, it is vital that some wayfarers in this life have doubts, and thus enable the division of epistemic labour.

On Lutterell’s account, this cannot be true, because the blessed one doesn’t entertain propositions in his or her mind, so the parts of former propositions could not be transported into the newly formed simple cognition. But Ockham encountered smarter opponents than Lutterell. To be sure, it’s perfectly possible to grant that the blessed one can think discursively, and still deny that wayfarer and blessed one can entertain the very same propositions.

Ockham’s opponent and fellow Franciscan Walter Chatton attacked his doctrines in many respects. Chatton granted that the blessed ones can think discursively, but he still denied that articles of faith can be demonstrated, on the grounds that such an article cannot be the conclusion of a heavenly demonstration. In this life we have premises from which we might infer a sentence and believe it; in heaven we have other premises from which we might infer a similar sentence, but – says Chatton – although our common manner of speaking might suggest that we are dealing with the same sentence, we have in fact sentences of a different kind. Because different kinds of premises acquired under different epistemic conditions entail different kinds of conclusions. And the blessed one could not but immediately judge that the wayfarer’s sentence is true. So Chatton infers that, although the knowledge of the wayfarer and the blessed one might be discursively structured, it still is of different kinds and thus incompatible.

  1. Ockham’s reply and Wodeham’s defence

Ockham concedes that the wayfarer’s premises are different from the premises available to the blessed one, but he denies that Chatton’s inference, according to which ‘different kinds of premises entail different kinds of conclusions’ is valid. As Aristotle had already established in his Posterior Analytics, the knowledge of the conclusion is not necessarily caused through the knowledge of the premises. Accordingly, Ockham claims that the blessed one might have the conclusion before he or she turns to assess the premises in order to prove the conclusion. And this conclusion may well be the article of faith that the now blessed one formerly entertained as a wayfarer in this life under insufficient epistemic conditions.

As I have said in the beginning, Ockham’s pupil Adam Wodeham defended this position. He counters Chatton’s tenets by pointing out that it does not contradict the beatific state of the blessed one if he or she forms the sentence of a wayfarer. This was exactly the idea that Chatton meant to attack, when he claimed that the article of faith of a wayfarer could by no means function as the conclusion of a demonstration, otherwise – Chatton had argued – we couldn’t say that the blessed one has more evident knowledge than the wayfarer.

Wodeham’s move against Chatton is even more subtle than Ockham’s initial defence. Wodeham repeats Chatton’s thesis that “the blessed one could not but immediately judge that the wayfarer’s sentence is true” and asks how the blessed one could achieve this judgement without engaging in a syllogistic discourse. Wodeham then explains that, if the blessed one assesses the sentence of the wayfarer, the blessed one has to demonstrate not his own but the wayfarers’ conclusion. Wodeham thus points to the logical blunder committed by Chatton: If the blessed one had solely knowledge of a different type, incompatible with the wayfarer’s knowledge, then the blessed one couldn’t know whether the wayfarer’s sentence is true. According to Wodeham, Chatton missed his target and refuted himself.

  1. Conclusion

Ockham and Wodeham argued against their opponents that the degree of certainty of knowledge and belief depends on the epistemic conditions of the knower. The difference between the knowledge of a wayfarer and of a blessed one is not a difference between types of knowledge but a difference of epistemic situations: If, for instance, I actually see something, I can form a reliable judgement about its existence and the properties belonging to the object seen. If I see God, then I can claim to have evident knowledge about him; in this life however, the sentences I form are expressions of faith. There is thus no difference with regard to the type of rationality, such as the difference between simple intuitions and complex propositions. All knowledge is processed in mental sentences which have a compositional structure and transportable parts.

Ockham’s point was not to raise hopes that we might meet the odd blessed one and ask him or her to demonstrate our conclusions. His thought experiment was designed to show that we need to assume a generally consistent structure of knowledge, if we want to claim compatible criteria for identifying and assessing judgements formed under different epistemic conditions. The scope of rationality that Ockham assumes comprises the knowledge of all thinking creatures, including humans, blessed ones and angels, creatures whose knowledge is shaped by common structures, rendering knowledge identifiable under different epistemic conditions. But for the idea of a common rationality to work, there also needs to be a division of epistemic labour. In this sense, Ockham defends a social notion of rationality.

  1. An afterthought

Ockham’s and particularly Wodeham’s subtle defence against those who assumed a plurality of incompatible types of lower and higher knowledge rest on interesting arguments. But as in many construals of logical form as not merely an instrument but the actual shape of knowledge itself, there remain troubling questions, one of which I’d like to sketch now at the end of this post: Do not many of these defences amount to transcendental arguments by means of which we proceed from supposed facts to the necessary conditions of their possibility? I think we are indeed inclined to infer from the fact that our reasonings are logically reconstructible, that it is a necessary condition of the possibility of such a reconstruction that the logical form really is a feature of our knowledge. Recall: Wodeham claimed that, for the blessed one to assess the wayfarer’s sentence, he or she needs to entertain and demonstrate the wayfarer’s sentence, and that hence their shapes of rationality have to be compatible. That seems quite conclusive.

If I assess a sentence S, the necessary condition of the possibility of my assessing S is obviously that I can form and entertain S, and thus that my mental capacities and structures have to be of a kind allowing for the entertainment of S. We couldn’t even talk about the blessed one doing so without implicitly or explicitly ascribing to him or her that capacity.

The trouble with such arguments is that our ascription of such capacities can be both indispensable and false, false at least in the sense that the capacity of entertaining S might completely exhaust the blessed one’s structured rationality, and that every other piece of his knowledge is indeed couched in a different, namely structureless type of rationality. On a lighter note we might urge this point as follows: When you have finished reading this post and you’ll start forming critical responses in your mind, you’ll certainly do so on the assumption that I have written this post in English and that we share a common rationality. You’ll say: the fact that Martin has written this post and might even respond to my comments rests on the condition that he has the capacity to speak English. But maybe I have just shot my bolt and wouldn’t know how to write anything else that might look like English; perhaps this post is just a sort of gap in my otherwise entirely different way of thinking. To put it in terms of the illustration I gave in the beginning: I might not be processing words, but simply be tossing out a sort of photocopy. So if I can’t reply to your comments the reason may well be that I didn’t write this post in English in the first place. The letters I have written down here may just look very similar to English words. But then again: how could I tell you that without sharing or anticipating your assumptions?


*This blog post is the revised version of an unpublished paper, first presented at the conference Dialectic on Trial, Cambridge, Clare Hall, 2005, and it contains some passages from my papers published here and here.


Ockham on the division of linguistic labour

One fascinating feature of language is that it enables us to talk about and thus handle things that we have no clue about. Although I have never seen a dinosaur, I can talk about dinosaurs and even recognise images of them. Arguably, this feature of language is owing to what Putnam called the “division of linguistic labour”. Although I don’t know anything about or at least not much about certain things, I can apply the words correctly. The reason I can do so is that I defer to other speakers, in this case experts who do have the knowledge I lack. According to this view, language is not a solitary tool like a hammer, but a social tool, only working in a collective, like a steamship. Although the social understanding of language is often portrayed as a modern idea, I’d like to suggest that it dates back at least to the early 14th century.

Thinking about language in wake of Putnam, we might find the thesis of the division of linguistic labour quite intuitive. But it doesn’t come naturally if we follow the standard interpretation of what is now known as the semantic triangle as sketched in Aristotle’s Peri hermenias (16a3-9). Here, Aristotle suggests that words relate to things in virtue of the mental concepts we have of things. The reason that words like “apple”, “pomum” or “Apfel” signify the same things is that they signify the same non-linguistic concepts. A common way to take this mediation of the word-thing relation through concepts is to assume that speakers must know the things they refer to. Using the word “apple”, then, requires me to have the concept of apple and thus knowledge of apples. – Now, in the first chapter of his Summa logicae, Ockham rejects this assumption. As is often the case, he frames his rejection as the proper interpretation of Aristotle. The details are a bit technical, but his own view is clear enough. In the translation by Paul Spade (p. 5-6), Ockham argues as follows:

Now I say that utterances are signs subordinated to concepts …, not because, taking the word ‘signs’ in a proper sense, these utterances always signify those concepts of the soul primarily and properly, but rather because utterances are imposed to signify the same things that are signified by the concepts of the mind, so that the concept primarily signifies something naturally, and the utterance secondarily signifies the same thing, to such an extent that once an utterance is instituted to signify something signified by a concept in the mind, if that concept were to change its significate, the utterance itself would by that fact, without any new institution, change its significate. The Philosopher says as much when he says that utterances are “the marks of the passions that are in the soul”. Boethius too means the same thing when he says that utterances “signify” concepts.

According to Ockham, then, words (translated as “utterances”) don’t signify things by signifying concepts first; rather it is enough that words signify the same things that are signified by the concepts that the words are subordinated to. Let’s compare these views schematically (the arrow indicates a signification relation):

Aristotle:            word     –>         concept  –>     thing

Ockham:            concept –>       thing

                            word      –>         “

So while the Aristotelian model suggests that each successful use of a word requires the user to have a concept that is signified by the word, Ockham’s model only requires that the word be subordinated to a concept. It is neither required that every actual use is mediated by a concept, nor is it required that the actual user does have the concept in question. Arguably, it is enough if some users have such a concept. If this is correct, we can assume that even those users who don’t have the pertinent concepts can use words correctly.

Although Ockham does not develop this idea straightforwardly into a thesis of division of linguistic labour, his rejection of the standard Aristotelian model clearly opens the door to such a view. Although she does not ascribe such a division thesis to Ockham, Sonja Schierbaum’s discussion of Ockham provides further evidence for such a reading. Ockham doesn’t develop this subordination semantics out of the blue. The discussion about the naming of God already led others before him to acknowledge that the use of words allows for talk about things that we don’t understand or have concepts of. (See here for more on this issue.) What Ockham adds in the Summa logicae is the idea that words can generally be used for referring to things irrespective of our concepts or knowledge. All that we need is that some other users have those concepts.

The upshot is (1) that conventional language is a more powerful tool than the mental concepts to which conventional language is subordinated and (2) that individual language use requires other users, especially for cases in which we lack pertinent concepts or knowledge.

The conventional signification of Lockean complex ideas*

In the course of writing up a paper on Locke’s theory of ideas for a volume on The Lockean Mind, it dawned on me that all complex ideas are conventional signs. So in what follows, I’d like to suggest that, while simple ideas are natural signs of their causes, complex ideas, even complex ideas of substances, are conventional signs that depend on linguistic consolidation.

In recent years, work on representation in Locke has reached some sort of consensus. Most scholars agree that simple ideas represent their causes. But it’s not at all clear what complex ideas represent. If substances are not given in experience, what do complex ideas represent? Do they have corresponding objects? Given Locke’s metaphysics, it seems safe to say that there are no apples, dogs, cats, trees or stones in the world, at least not properly speaking. Let’s unpack this a bit: While the ingredients of complex ideas, i.e. simple ideas, represent the qualities (= properties) that cause them, the resulting complex idea might be said to represent its cause insofar as the simple ideas represent their causes. So the complex idea of an apple might be said to be a natural sign of its cause insofar as the simple ideas (of which it consists), say the ideas of round, green, sweet and firm, are natural signs of qualities. Yet, this cannot be the last word, because complex substance ideas seem to have no direct correlate that they represent. Arguably, we cannot say that the things we think of in virtue of complex ideas are given in experience. My idea of an apple is not simply caused by an apple. Rather the complex idea of apple is made up of ideas of qualities, while the apple as a substantial bearer of the perceived qualities is not itself given in experience but presupposed. Looking at the causal explanation of such ideas, we cannot simply invoke extramental qualities. We also need to take into account the tacit mental operations that form the complex idea out of simple ideas (Essay II, xxiii, 1).

What, then, do complex ideas represent? As I have just said, the idea of an apple can of course be analysed into ideas of qualities. But if the apple idea as such is not taken as a sign of the causes (qualities), what does it represent? To answer this question, it is important to note that Locke appeals to an old distinction between two kinds of ideas, archetypes and copies or ektypes. Traditionally, archetypes would be taken to be ideas in the mind of God, where they serve as blueprints of things to be created. Locke doesn’t appeal to divine ideas, but he still makes use of this distinction when discussing adequate and inadequate ideas (see Essay II, xxxi). Arguably, Locke assumes that our mind forms, stores and names such archetypes when we encounter things or invent moral categories. The archetypes are what Locke calls nominal essences in book III. What does this mean? When we have an occurrent idea of an apple, this idea can be seen as a copy of the archetype through which we recognise the thing as an apple. So when we see an apple, we have in fact an idea that matches to some extent the archetype of apples in our mind.

This archetype-copy relation also forms the foundation for Locke’s famous criticism of our assumption that we have cognitive access to real essences of things. According to Locke, we tend to assume that a present idea of an apple matches the real essence of an apple. But this is a mistake, argues Locke. In fact, we refer it to the nominal essence, which is nothing but an abstract idea with a name annexed to it (see Essay III, iii, 15).

An immediate worry arising from this account is that such archetypes would be highly subjective or instable. After all, nominal essences (or archetypes) of substances are formed in our various accidental encounters with things. But if we invoke Locke’s discussion of language, we can see that he meets this worry. Briefly put, the fact that nominal essences are annexed to names means that these essence ideas are socially consolidated. The linguistic community a speaker is part of will sanction some uses and confirm other uses. This way, the abstract ideas are stabilised and even rectifiable in the light of new discoveries. So if I happen to call apples tomatoes, other users will correct and amend my use of the term, and thus make me adapt the conventional nominal essence.

The upshot is that complex ideas of substances (and mixed modes) are conventional signs to some degree. Although complex ideas consist of simple ideas that can be seen as natural signs of their causes, the complex ideas ultimately function as conventional signs for us. What they present to our mind is not solely determined by what causes their ingredients but also by the patterns that are consolidated within a linguistic community. This means that my occurent idea of an apple, while caused by certain qualities, presents to my mind an apple that I can recognise as an instance of what we call apple, because it has to match the socially consolidated criteria of the nominal essence of apple. Were it not for the historically grown categories salient in my speech community, for instance those for certain kinds of fruit, it is possible that my mind would register the recurrent ideas of certain qualities without ever recognising them as pertaining to a certain kind of thing.

Locke makes this point clear when he writes that ideas without names annexed to them would resemble a library with unbound books in which we could access pages without it being determined to which books the pages belong: ‘He that has complex Ideas, without particular names for them, would be in no better Case than a Bookseller, who had in his Ware-house Volumes, that lay there unbound, and without Titles …’ (Essay III, x, 26-27). The point is that the precise ingredients of our complex ideas, the structure and amount of simple ideas belonging to them, are determined by the naming practice in a given speech community. Thus, although the ingredients themselves have a causal origin, the way the ingredients are bound into bundles and used in thinking is a matter of convention.


* This piece is now kindly featured by Locke Studies 19 (2019).

“How would you arrange the deportation of my father?” On responsible (free) speech. A response to Silvia Mazzini

Could you tell me, face to face or in writing, how you would go about having my father deported? – Why, you ask? – Well, maybe you think he is a burden for society. After all, he is quite old by now. So how do you do get it arranged? Should some people be sent to fetch him? Perhaps at night? Go on, then! –

You, gentle reader, probably don’t have such desires. But if I follow the political discussions in the Netherlands and other countries, many people want that. Only they don’t tell me personally; they talk about certain groups, not to me.

Ah, it’s not old people, you say, just Muslims? So they don’t want to come for my father? Well, lucky me then… Should it make a difference whether people want to deport my or someone else’s father? Well, it makes a difference, but does it matter? Not much. –– The point I would like to suggest is that we can imagine that certain opinions concern us directly, even if they don’t. In a controversial discussion between two opponents, such imaginations can help both interlocutors to make the conversation more personal, concrete, emotional and thus responsible. Following up on my last post, I would like to develop some ideas, then, how we can turn free speech into responsible speech.

In my last post, I tried to show that our disagreements about the limits of free speech are owing to two different ways of understanding how language works. Ultimately, I suggested that the crucial limit of free speech should be determined by the responsibilities we have as speakers. But I didn’t say much about these responsibilities themselves. Commenting on the post, my colleague Silvia Mazzini suggested that responsibility could be seen as offering the other the ability to respond:

Maybe we could then interpret [responsible freedom of speech] like Levinas did: responsibility is the “ability to respond”. In this sense, freedom of speech would mean that all the people involved in a dialogue are able to respond – that they have the intention to consider the different positions of the others.

This strikes me as the way to go. What I like about this idea in particular is that it doesn’t require us to provide a complicated catalogue of virtues or rules. Rather, the responsibility is imposed through the very fact that the opinion is not voiced as a statement about others but to others.

What’s the big deal, you might ask, does it really make such a difference whether I offer my opinion about a policy regarding a group of people to someone in particular? David Livingstone Smith’s work on dehumanisation made me see one point in particular recurring again and again: Although it might be simple to imagine doing harm to a certain group in the abstract, it is really hard to do something harmful to someone directly in front of you. (That is why dehumanising tactics are employed: it is easier to harm someone if you think of them as not really human.) Arguably, this carries over to speech acts. My hunch is that it is much harder to direct hate speech at someone in particular (rather than speak abstractly about members of a group).

My idea is, then, that it is easier to act as a responsible speaker, if you are addressing someone in particular directly. There are a number of reasons for this. Interacting with a concrete person, we are more likely to respond with adequate emotions and empathy, and we have to face the response. Although a face-to-face encounter will be best, I think this will even work in online communication. It makes a difference for me as a writer whether I imagine you, whoever you are, as a concrete person who might frown or agree. Or whether I simply toss out statements about abstract ideas, however much they might affect you. The point is, thus, that we shouldn’t always try to amend online debates by being as rational as possible or by cancelling out emotions. Rather, the task would be to facilitate adequate social emotions necessary for responsible interaction. Addressing others directly should have two consequences: (a) it should be more difficult to objectify and thus to harm the interlocutor; (b) it should invite the other to respond and make me anticipate some response. Thus, if we get people who utter opinions to address people directly in this way, they will speak more responsibly, rendering free speech not a battleground but a possibility for genuine and considerate exchange.

So far, so good. Of course you might have objections, but my worry at this point is not how to justify my idea. Rather, I see the main challenge in implementing it. I think we should give it a try and then see how well it works. So how can we change our conventions? How can we get from talking about people to taking to them? This is an open question, but at this moment I can think of four steps in the relevant contexts:

  • Change speech acts from third-person to second-person sentences: Saying that you should leave this country is much harder than saying that blog readers should leave this country. I’d think twice about what’s going to happen if i did so.
  • People can stand in for targeted people: If you hear someone going on about a religious group, you can respond as if you were targeted. The point is not to lie, but to offer yourself as a possible interlocutor (which might be more effective than just saying that the speaker is a bad person).
  • As a possible interlocutor you can demand the other to (empathetically) imagine your situation: It might make a difference to ask your opponent how she thinks the deportation of your father should be arranged. Rather than discussing the rights and wrongs in the abstract.
  • Dehumanising language must be rejected. Of course, there are limits. It is vital to state that, if your interlocutor crosses a red line.

Now you might think that all of this is too difficult. I doubt that. In the face of what we often call political correctness, we have acquired a lot of vocabulary and changed some of our speaking habits. Now we can adjust our imagination and syntax a bit. Of course, this will take time. But I really hope that you and I as well as (other) people in education, in companies, in the press, moderators in the media, citizens in online or analogue discussions gradually train and learn to adjust their language and address people directly. Yes, it will be harder to offer your opinion, but it will also be more fruitful. – At this point, I’m suggesting this and hope for more ideas about means and ways of implementing it. Ideally, we’ll find that this or something like it turns out to be a viable way of amending political discourse.

By the way, this should cut across the entire political spectrum. It has, for instance, become fashionable to engange in what is sometimes called leftist populism or target the group of “old white males”. Whatever your contention might be, if you want to tell someone like my father that inverse racism isn’t a thing, you won’t get him to respond sensibly if you target him as a member of that group. We act through language. And the way we act in our words is palpable, it affects individuals, and individual people are likely to respond in kind. Verbal attacks affect us, irrespective of the side we think we are on. Thus, whenever we want to make a point that affects others, we should try and address them directly. Conversely, if we encounter problematic opinions, we don’t need to shut them down. To respond on behalf of a targeted addressee, as if you were addressed directly, might be more fruitful in maintaing adequate standards and emotions.

Finally, it goes without saying that I am worst at following my own advice. So please don’t call me out too harshly.