Do you value topic-continuity? The moral foundations of Cappelen’s insistence on ‘topic-continuity’ and reasons for resisting them

ABSTRACT

The article reveals the pragmatic implications of Herman Cappelen’s account of ‘topics’ in his contribution to the conceptual engineering literature. I show that Cappelen’s introduction of the category of ‘topics’ serves the pragmatic goal of having a convenient handle to account for ‘continuity in revision’, and that his general insistence on ‘continuity’ is motivated morally and strategically. In asking what accounts for continuity, Cappelen’s ‘topics’ are not defined by content or any other fixed set of rules or criteria. Topics are metaphysically lightweight and defined pragmatically and as we go: speakers talk about the same topic when we (and they) attribute that they do. But why should we do that? Why should we aim for continuity and why should we think it is possible in general? I contrast Cappelen’s insistence on continuity with Rorty’s appeal for discontinuity, and trace both of their positions back to their respective moral background assumptions and their assumptions about what communication is (and is for) and, in turn, what philosophy is (and is about). Further, I question the role the ‘continuity’ claim plays in the current redefinition of linguistic philosophy after the ‘death of the linguistic turn’.

KEYWORDS:

• Conceptual engineering
• metaphilosophy
• neopragmatism
• Rorty
• disruption

1. Introduction

In Fixing language (2018), Herman Cappelen introduces a category into the conceptual engineering (CE) literature that has attracted a lot of attention: ‘topics’. What the category does is account for what I call ‘continuity in revision’, or what others have called ‘conceptual engineering without conceptual change’ (Jackman 2020). The attribution of the notion ‘sameness of topic’ expresses that after some engineering process is concluded, we have not simply changed the subject but are still talking about the same thing – or topic – the talking about which we have set out to improve in the first place.¹ But why is this important? And with respect to which semantic categories could we ever decide the question?

The discussion of Cappelen’s ‘topics’ has taken many forms, with positions arguing both for and against the category as such, about how to determine continuity or not, and a variety of other connected issues, such as the notion of ‘functions’ (for some recent accounts see Belleri 2021; Jorem 2022; Knoll 2020; Koch 2021; Nado 2021; Riggs 2021; Thomasson 2020). This article is not a contribution to this fine-grained discussion within CE but instead a reflection on how the discussion regarding topics and their continuity or disruption allows us to restate an older question: namely, what the linguistic turn entails or should entail.

Cappelen himself thinks that equating the linguistic turn with descriptivism is a ‘historical aberration’ (Cappelen 2017, 758). For Cappelen, philosophy of language is, and always was, in the business of ‘normative reflections on language’ (Cappelen 2017, 755). Obviously, everything hangs on what one means by ‘norms’ here, and whether what analytic philosophy of language does is ‘discover’ these norms or if it is in the business of ‘making’ them. My view is that Cappelen’s ‘topics’ are a way to dismiss the infelicitous distinction between ‘making’ and ‘finding’ by guaranteeing flexibility with respect to deciding between continuity or disruption of meaning.² ‘Topics’ allow us to make sense of the fact that we do in fact cooperate via using language, even though sometimes we do not all mean the same when using the same word-string, or terms, in the same context.

Once the notion of ‘topics’ – which allows for changed meanings to still address the same subject, issue, or ‘topic’ – is in place, one can then ask a question regarding sameness of topic. However, as Amie Thomasson puts it:

The deepest, though not most direct, response to the generalized challenge³ is to urge that we not presume that there is an objectively correct ‘discovery’ of what does/does not count as sameness of topic, concept, or term. What we count as sameness of concept or term may aptly be engineered or negotiated differently depending on the purposes we have. (Thomasson 2020, 442)

What I want to do in this article is: (A) show that Cappelen’s approach to topics is compatible with the pragmatist view defended by Thomasson; (B) distinguish between the notions of ‘topics’ and ‘topic-continuity/disruption’ and show that once one has embraced (A), Cappelen’s further insistence on topic-continuity – beyond the mere introduction of ‘topics’ as a technical fix to capture ‘a certain degree of continuity between changed meanings’ – is informed by his moral background-assumptions about what communication is and should be for and how this connects to his take on the linguistic turn; (C) suggest what further moral ends could be served by emphasizing rupture instead of continuity, and (D) end with a plea for flexibility between continuity and rupture.

To meet these aims I proceed as follows: I will first briefly reconstruct Cappelen’s account of ‘topics’ and show why his account is compatible with pragmatism (Section 2); secondly, I will explain why and how Sally Haslanger and Rachel Sterken argue for continuity and rupture of topics, concepts, or word-meaning-pairs in contexts of social justice (Section 3); third, I will show that the same choice between continuity and rupture which Haslanger and Sterken discuss in the political context applies in the meta-philosophical context to which Cappelen is contributing, and investigate why he opts for continuity (Section 4); then I will contrast Cappelen’s moral background assumptions with respect to meta-philosophy with Richard Rorty’s moral background assumptions, and give a rationale for why and when to endorse, or at least be open to, topic-disruption instead (Section 5); finally, I will spell out what all of this means with respect to the legacy of the linguistic turn and the role that ‘conceptual engineering’ plays or wants to play in this story (Section 6).

2. Cappelen’s ‘topics’ and his further semantic and metasemantic commitments

The reason Cappelen’s account of topics is compatible with pragmatism is a combination of Cappelen’s general semantic and meta-semantic commitments and Cappelen’s definition of ‘topics’ (2018).

Cappelen’s semantic and meta-semantic commitments entail, among other things, that there are (contra Chalmers) no ‘bedrock-concepts’ (Cappelen 2018, 192–195) and there are (contra Eklund) no normative limits of the sort that, for instance, would say that the concepts we use for assessing conceptual change cannot themselves be up for grabs (195–198). For Cappelen, conceptual engineering knows ‘no natural endpoint’ (194), goes ‘all the way down’ (195), and also encompasses the meta-semantic level because ‘there’s no good reason to assume that only the semantics is in flux. We should also expect the metasemantics to be in flux’ (69). As Cappelen puts it: conceptual engineers of his ilk

lack any solid resting point: we representational skeptics need a language in order to engage in critical reflection about representations. But what about that language? Shouldn’t we be critical about that as well? The answer is ‘yes’: we should be skeptics throughout, and this makes our project exhausting. (6)

I take these commitments to be in line with pragmatism, insofar as Cappelen’s endorsement of ‘contestation throughout’ (8), where even ‘success conditions are up for grabs’ (8), and where there are no final answers to be expected but only ‘progress reports’ (6), resembles positions held by Rorty. Cappelen himself draws the analogy with respect to metasemantics.⁴ He cites an early piece of Rorty’s on metaphilosophy, where Rorty states that

philosophy is the greatest game of all precisely because it is the game of ‘changing the rules.’ This game can be won by attending to the patterns by which these rules are changed, and formulating rules in terms of which to judge changes of rules. Those who take this view hold that philosophy in the old style – philosophy as ‘metaphysics, epistemology, and axiology’ – needs to be replaced by metaphilosophy. Members of this school are, as it were, the metaphilosopher’s metaphilosophers: since any metaphysical, epistemological, or axiological arguments can be defeated by redefinition, nothing remains but to make a virtue of necessity and to study this process of redefinition itself. (Rorty 1961, 301; Cappelen 2018, 69, 153)

Cappelen explains that Rorty, obviously, did not endorse ‘my theory of conceptual engineering – he went on to use these points for somewhat tangential purposes’ (Cappelen 2018, 153). I think Cappelen is right. In fact, Bjørn Ramberg and I cite the same Rorty piece to illustrate what Cappelen calls Rorty’s ‘tangential purposes’: on our reading, Rorty after a certain point went on to engage in redefinition instead of only studying it (Huetter-Almerigi and Ramberg 2022). He was trying to do some conceptual engineering inside the philosophy departments of his time, one could say now, but with little success. I will come back to this point.

So much for Cappelen’s general commitments. Cappelen dedicates many pages of Fixing Language to his account of ‘topics’ and to motivating their relevance. The following succinct definitions are key to my purposes:

topics are more coarse-grained than extensions and intensions, and so expressions that differ with respect to extensions and intensions can be about the same topic. (2018, 101)

This creates the conceptual space needed to account for ‘continuity in revision’:

Topic continuity is compatible with changes in extension and intension – the semantic values of ‘F’ can change, whilst we continue to talk about F. (2018, 54)

When ‘F’ and ‘F*’ are still addressing the same topic, the moves from ‘F’ to ‘F*’ can qualify as improvement⁵ (2018, 148ff). Particularly important for my analysis here is that

the constraints on topic continuity are not fixed, but essentially contested. (2018, 54; my emphasis)

So, whether ‘F’ and ‘F*’ still pertain to the same topic is a question that cannot be answered by reference to the norms and rules of pure semantics. Whether there is topic-continuity or not is, from a semantic point of view, essentially up for grabs. Therefore, we need to look for further reasons for why we should construe contributions as continuous with or disrupting a certain topic that are not exhausted by the norms and rules of the semantic game.

The reasons that Cappelen gives for insisting on topics and topic-continuity are that they allow us to account for the ‘continuity of inquiry,’ they can help handle verbal disputes, and they allow us to account for disquotational reports over time (2018, 100–104). All of Section 4 below is dedicated to Cappelen’s insistence on ‘continuity of inquiry’ which, I will suggest, is not motivated semantically – or even more generally theoretically – but rather morally. To make that point below, let me stress here that ‘topics’ come with no further metaphysical strings attached:

I don’t have a metaphysics of topics and their identity conditions over time. Instead, I have an account (or a description) of the contestation over when it’s legitimate to say, ‘They’re still talking about (or discussing, describing, or …) marriage (or freedom or …).’ The model gives the conditions under which we describe two people as samesayers: they can be samesayers even when the semantic values of their words diverge […], but what makes that true isn’t a new semantic value that is identical between the two of them. (2018, 140)

‘Topics’ are metaphysically lightweight and, to avoid infinite regress (2018, 141), don’t refer to a new level of content:

Sharing topics, as I construe it, is not the sharing of a new level of contents (a content neither of the two groups expressed semantically). For there to be topic continuity is for the groups to be treated as samesayers for certain purposes. So construed, the foundation is ‘ineffable’ in that it’s not a new saying or a new content. (2018, 198)

To sum up: When investigating what accounts for topic-continuity, Cappelen does not suggest a new level of content. Instead, topics are metaphysically lightweight and defined pragmatically as we go: ‘What creates unity in topic are acts of treating them [the speakers] as samesayers’ (2018, 198; my emphasis). Speakers talk about the same topic when we (or they) attribute that they do.⁶

But why should we do that? Why should we aim for (or state) continuity if there is no theoretic necessity of doing so? What further reasons could one have to insist on continuity? In Section 4 below I will investigate Cappelen’s commitment to the ‘continuity of thought, talk, communication and inquiry’ (2018, 102) as a practical reason for insisting on continuity, and show that his commitment is motivated both strategically and morally. But let’s look first at another context where the choice between continuity and rupture of topics is motivated by reference to practical reasons and consequences.

3. Political-operational reasons for insisting on continuity or disruption of topics

There can be political or societal reasons for insisting on continuity. Consider, for example, the answer Sally Haslanger gives for not changing the subject⁷: because ‘Sharing concepts […] is crucial for communication and coordination’ (Haslanger 2020b, 238). Haslanger’s insistence on amelioration instead of replacement⁸ is informed by considerations of the political effectiveness of our utterances. The sharing of social meaning⁹ has an impact on the coordination of our actions.¹⁰ By insisting on the continuity of concepts like ‘marriage’ or ‘parent’, while nonetheless making changes to their meaning, overall communication is maintained and secured. This way, so goes the hope and theory, we can ameliorate certain practices without risking problems for the overall coordination and smoothness of our social exchanges (Haslanger 2012, 381–403, 2020a). Therefore, the reasons for continuity lie in securing social peace and justice by considering what role linguistic uptake is playing in these areas. Haslanger’s strategy in a catchphrase could be: the more we go for uptake,¹¹ the smoother the transition will be, which, in turn, renders the transition more likely.

However, in the same context Rachel Sterken has argued that in certain cases rupture and friction might have just as fruitful effects (Sterken 2020). In what she calls ‘transformative communicative disruption’ (2020, 418), communication is not smooth but intentionally interrupted. Sterken argues that in certain cases it might be better to ‘break communicative chains’ (2020, 421), or to baptize new word-meaning pairs, ‘while, simultaneously, attempting to render defective the interpretative common ground of the original word-meaning-pair’ (2020, 421). This creates a shock effect that prompts speakers to reflect on their linguistic practices (2020, 430). The cost of this intervention is that it leads inevitably to a transition period of misunderstanding and confusion (2020, 426–428). However, according to Sterken, overall ‘It’s good that changing language leads to miscommunication and confusion, because that can cause speakers to reflect on their language, and that will lead them to focus on its flaws and ways to improve them’ (2020, 433). Here, impact is achieved by actively dismantling, instead of securing, the common ground one wants to substitute.

What is important for my purposes here is that both proposals, Haslanger’s and Sterken’s, argue in a strategic or practical register rather than by claims to theoretical necessities – a point that Cappelen also acknowledges. Cappelen writes that for Sterken ‘miscommunication itself is part of the larger communicative strategy’ (2018, 133), but he thinks that ‘to push changes in meaning in a way that does not preserve topic is risky’ (2018, 132). Therefore, it’s the prospective practical consequences that render the individual moves aspirational or risky, where the practical consequences are calculated on the basis of conjectures about what happens when going for either continuity or rupture.

4. Cappelen’s reasons for insisting on continuity

I will now argue that Cappelen’s insistence on topic-continuity serves the same purpose in philosophy of language and metaphilosophy that it serves at the political-societal level for Haslanger and Sterken: it is informed by Cappelen’s further commitments, and it is modeled on his conjectures about the best way to secure impact for ameliorative or engineering projects.

We saw above that, as there are no fixed semantic or meta-semantic norms in Cappelen’s framework that allow us to answer the question about whether to insist on continuity or rupture of topics, one has to give further, non-theoretical reasons for insisting on one or the other. The non-theoretical reason that Cappelen gives is that an insistence on continuity of topics is needed to secure the ‘continuity of thought, talk, communication and inquiry’ (2018, 102). But what exactly does that mean?

I believe that ‘continuity’ or ‘unity of inquiry’ comes in two forms in Cappelen’s book:

• – ‘unity of inquiry’ as an implementation strategy

• – ‘unity of inquiry’ as a moral-ethical commitment

(Though I don’t believe there needs to be a clear-cut distinction between the two).

To understand what I mean by ‘“unity of inquiry” as an implementation strategy’, one needs to take a closer look at Cappelen’s take on lexical effects (2018, 122–134). Lexical effects are all the effects that words have which cannot be explained by our given semantic and pragmatic taxonomic repertoire. Cappelen gives various examples of lexical effects, amongst which proper names (e.g. no one should name their child ‘Hitler’) and brand-names (like ‘Coca-Cola’), and the effects caused by metaphors (à la Davidson¹² and Lakoff) (2018, 122–134). The idea is that

The change in lexical item would change […] [people’s] behavior. More generally, one lesson from reflecting on brand names and the theory of brand names (something philosophers of language and linguists should do much more) is that lexical effects are immense. The lexical items themselves have broad cognitive and emotive effects which are not captured simply by talking about Gricean implicatures, presuppositions, or other standard pragmatic phenomena – in part because they are not entirely, or even for the most part, cognitive effects (they don’t add another level of content). (2018, 124)

There is no comprehensive theory of what lexical effects exactly are or how to trigger them, but this does not render the category any less important. In fact, they are especially important for CE-enterprises. Cappelen writes:

I’ve given […] some illustrations of lexical effects, but not a general theory. I don’t have one. I think they are a hugely varied category. They can inspire, trigger associations of various kinds, make you happy, cheerful, motivated, sad, angry, etc. They can certainly affect your dispositions to behave. This is a point made particularly vivid by brand names, but if a name like ‘Coca Cola’ can make you more likely to spend money on a bottle of liquid, then most likely a name can also make you more disposed to accept, say, a philosophical theory (it is, after all, cheaper to accept a philosophical theory than to buy a bottle of Coke, so we should expect the barriers to be lower). (2018, 130)

As for the language that philosophical texts are written in, ‘It might turn out that even for us [philosophers], a significant amount of our work consists in trying to trigger the right kinds of lexical effects’ (2018, 129). This is the radical self-application that Cappelen also embraces in other places, for instance when explaining that the corollaries that come with being a ‘representational skeptic’ apply also to the language in which this very sentence is written.

Turning to philosophy (though the point here has nothing specifically to do with philosophy), I suspect that the choice of whether to use terms such as ‘reductionist’, ‘feminist’, ‘experimental philosophy’, ‘intuitive’, ‘anti-realist’, ‘relativist’, ‘analytic’, etc. is often in large part guided by the non-semantic and non-pragmatic lexical effects the use of these terms will have. Roughly the idea is: show me which words you use, and I’ll tell you who your friends are (or who you want your friends to be). (2018, 128–129)

This extends to the terms ‘unity of inquiry’, ‘exchange of ideas’, ‘rational discourse’, and ‘genuine communication’. This is not to say that Cappelen is consciously and intentionally exploiting lexical effects. Cappelen shuns what he calls (intentional) ‘exploiters’ of lexical effects, and he shuns them for moral-ethical reasons, as we will see in a minute. Rather, I am saying that his text (2018) is apt to trigger these effects even beyond his intentions. This is just saying what Cappelen says in the citation above: consciously or not, and apart from embracing, or not, what we are saying (though I have no reason to doubt that Cappelen does embrace what he says), what we say, at the level of the exact word-strings we use, can have effects on our behavior and, by extension, on group-cohesion and the probability of successful implementation of certain changes. In short, we use the same words our friends use. Enacting unity by displaying our shared linguistic repertoire strengthens emotive bonds.¹³ There is an existential side to this which makes further investigation of lexical effects a crucial project. For Rorty, words can become an existential shelter (Penelas 2019; Llanera 2020). We feel at home and caressed by certain words, and repulsed and alienated by others. A dear friend of mine once told me that he instantly fell out of love with his ex-partner when they were watching TV together and he heard her call Evita Peron certain words. The words she used were a causal slap in the face to his love for her. He could not be with her anymore after that.¹⁴

Certainly, an analogous case can be made for the words of ‘unity of inquiry’, ‘exchange of ideas’, ‘rational discourse’, and ‘genuine communication’, when uttered in the context of analytic philosophy. People feel at home with these notions and then, perhaps, are more inclined to take in the more radical claim that there are no norms for guaranteeing these concepts or even for knowing for sure if and when we encounter them. Again, what I am saying is not that Cappelen is exploiting lexical effects. What I am saying is that if it were a strategy, it certainly is chosen well and that considerations regarding implementation and moral commitments need not come apart. Let me now come more directly to Cappelen’s moral commitments, i.e. his declared (practical) reason for insisting on topic-continuity, which is that topic-continuity secures the ‘continuity of thought, talk, communication and inquiry’ (2018, 102).

Beyond lexical effects¹⁵ (which might be intentional or not), the content of Cappelen’s book (2018) certainly is in continuity with large parts of the analytic tradition, a tradition which embraces values which Cappelen declaredly and intentionally also embraces, and which lie in not only using the words, but in – and this is what makes it moral – treasuring ‘continuity of inquiry’. Cappelen thinks that there are ‘Three varieties of conceptual engineers’ (2018, 148–154): Topic-improving engineers (like Haslanger on ‘gender’ [Haslanger 2012, 221–247]), semantic engineers (like Kevin Scharp on ‘truth’ [Scharp 2013]), and lexical-effects exploiting engineers. Cappelen stresses that this is a theoretical reconstruction and that ‘from a practical point of view, there isn’t always a sharp distinction’ (2018, 149). The three varieties can come in mixed forms and the choices aren’t always conscious. However, what distinguishes the lexical-effects exploiting engineers is that

Exploiters [who don’t care for topic-preservation] undermine rational discourse by encouraging verbal disputes and in so doing undermine continuity of inquiry. They treat speech as a medium of manipulation, not as a medium for communication (i.e. as a medium for the exchange of thoughts and ideas). There are of course Exploiters with good intentions, but the overall effect of their exploitation is to contribute to and encourage a use of language that undermines what we should treasure the most about it: the continuous exchange of ideas. Exploiters are in effect anti-intellectualist opportunists that contribute to a destruction of genuine communication. (2018, 133–134)

This is what I referred to above as “continuity of inquiry” as a moral commitment’: that what we should ‘treasure the most’, and what we should put a ‘high value’ on (2018, 49), is the ‘continuous exchange of ideas’, where these ideas stand in a line of continuity guaranteed by rationality (‘rational discourse’ [2018, 133]). In joint work with David Plunkett, Cappelen similarly describes this ethical dimension: ‘consider Carnap’s aim of modifying language to allow multiple people, from multiple places, to engage in collective rational inquiry. Making that possible was in part a political aim, tied to a democratic, enlightenment view of politics that ran through Carnap’s work’ (Cappelen and Plunkett 2020, 18). I call this view ethical instead of political because it points to an ethos which can be translated into political consequences if political measures are taken. The underlying conjecture seems to be that valuing continuity of inquiry implicitly promotes social peace.

What I want to do now, and in the rest of this article, is not to undermine this position but to complement it with another vision and possibility which, as I see it, is equally compatible with Cappelen’s semantic framework (in Section 2), and which is equally based on moral considerations and conjectures about peaceful convivence.

5. Rorty’s reasons for insisting on ruptures

As I said above, Rorty’s and Cappelen’s frameworks resemble one another in their radicality of what is up for grabs. What differs¹⁶ are their moral commitments and their ideas about how to best serve our aims via linguistic interventions; in short, whether to go for continuity or embrace disruption. From a Rortyan point of view, views similar to Cappelen’s insistence on the ‘continuity of inquiry’ are due to the influence that epistemology has had in shaping what we post-enlightenment people think communication is:

For epistemology, to be rational is to find the proper set of terms into which all the contributions should be translated if agreement is to become possible. For epistemology, conversation is implicit inquiry. (Rorty 1979, 318)

Famously, Rorty believes that this picture should be superseded by hermeneutics, and his reasons are equally of a moral nature. Whereas Cappelen emphasizes treasuring ‘unity of inquiry’ via topic-continuity¹⁷ – seen as underlying democratic utopia understood along the lines Cappelen and Plunkett attribute to Carnap (‘to allow multiple people, from multiple places, to engage in collective rational inquiry’ [2020, 18]) – Rorty thinks we should treasure what Ramberg and I have called ‘expansive solidarity’ Huetter-Almerigi and Ramberg 2022, 15). The latter is guaranteed by an increased resilience in contexts of incommensurability, because the idea that everyone’s claims and utterances are commensurable (or should or will or could one day be) is open to the violence now better known as testimonial and hermeneutic injustice (Fricker 2007).¹⁸ If we do not have an open ear for voices who are not commensurable with our own, who do not ‘engage with us’ in ‘collective inquiry’, and who maybe never will, then many social injustices will forever remain unheard and unaddressed. ‘Expansive solidarity’ aims to keep the hermeneutic door open for voices which are hard (or even up to impossible) to understand. Therefore, Rorty’s insistence on the possibility and omnipresence of topic-disruption is a complement to Cappelen’s view. To be clear, the difference is one of emphasis and not of kind – as I’ve stressed, the job-description of Cappelen’s notion of ‘topics’ is exactly to allow for discontinuity of meaning while offering a conceptual space (‘topics’) where, notwithstanding contestation over meaning, we still agree to pursue the same goals or aims.

Rorty thinks that to be better democrats and better moral human beings, we should train ourselves in the experience of not only not speaking the same language, but also the possibility that we will never share the same aims. The citation above goes on like this:

For epistemology, conversation is implicit inquiry. For hermeneutics, inquiry is routine conversation. Epistemology views the participants as united in what Oakeshott calls a universitas – a group united by mutual interests in achieving a common end. Hermeneutics views them as united in what he calls a societas – persons whose paths through life have fallen together, united by civility rather than by a common goal, much less by a common ground. (Rorty 1979, 318)

The lesson that Rorty takes from Kuhn is to abandon the ‘quest for commensuration’¹⁹ (Rorty 1979, 317) and to learn to live with ruptures – ruptures of meaning and of topics, to put it in Cappelen’s terms, though Rorty speaks only of meaning, which in his case comes with the flexibility of topics (see footnote 2). Hermeneutics, in this sense, is an exercise in enduring incommensurability:²⁰ it trains us to have an open ear for voices that come from beyond our current normative realm. These voices, which operate on norms (and goals) that we do not share – up to the potentially radical point of incommensurability – need not necessarily be the radical Others described by Spivak (1988), who famously argues that Indian subaltern women ‘cannot speak’ (Spivak 1988). What she means is that these women have no access to a shared linguistic community in which their voices could be heard or understood or even count as a voice. They are radically Other to us because there is no potential point of contact with their language game, up to the point that Spivak questions (much in line with, e.g. Davidson 1973) whether they even have a language.

The idea that this Other, in different guise, need not be an Indian subaltern women but could be our partner or our children is worth considering. Maybe to embrace the possibility that we will never bring them under the same norms or share goals (we will never speak the same language, understand them, or agree with them), that miscommunication is more likely to be the rule rather than the exception but that we need to get along anyhow, is as much an ethos that encourages peaceful cohabitation as Cappelen’s enlightenment ideal.

In Rorty’s (1961) piece on recent developments in metaphilosophy, which Cappelen cites to underline Rorty’s affiliation to something at least paralleling Cappelen’s definition of CE (‘philosophy is the greatest game of all precisely because it is the game of ‘changing the rules’’ [Rorty 1961, 301; Cappelen 2018, 69, 153]), Rorty goes on to ask himself what the goal of a game whose rules are ever-changing could be. He concludes that it can only be to ‘keep communication going’ (Rorty 1961, 302): ‘To keep communication going is to win the game’ (Rorty 1961, 302). Ramberg and I have pointed out that as to study the ‘process of redefinition as such’ (Rorty 1961, 301), which is what we do along the road of keeping conversation going, ‘approaches governed by different sets of rules are meant to be brought into communication, commensuration cannot be a necessary condition for the interpretative activity of the philosopher’ (Huetter-Almerigi and Ramberg 2022, 3). This seems to indicate a different sort of communication from Cappelen’s insistence on ‘rational discourse’ and ‘genuine communication’ (2018, 134, my emphasis) which points to the ‘unity of inquiry’ (although, here as well, the difference might be one of degree and emphasis, and not of kind). Certainly, Rorty’s ethos points to openings and for us to learn to handle, and live with, disruptions.

6. Redescribing or discontinuing the linguistic turn

Now let me come to what all of this can mean with respect to the legacy of the linguistic turn or for analytic philosophy of language as such. For Cappelen ‘The linguistic turn is dead’ (Cappelen 2017, 734), if what one means by the linguistic turn is the movement epitomized in Rorty’s collection The Linguistic Turn from 1967 (Rorty 1992b). In Cappelen’s words: ‘Something weird and unfortunate happened around 1970: much analytic philosophy of language took a sharp descriptivist turn and the various engineering projects were downplayed’ (2017, 758). What others might identify with the linguistic turn, i.e. the attention brought to language at the beginning of the twentieth century that gave birth to much of analytic philosophy (see e.g. Dummett 1996), for Cappelen should be better described as ‘Conceptual Engineering’, as most of these projects – namely Frege’s, Carnap’s, and other early positivist views – were engaged in normative instead of descriptive inquiries, and their occasional descriptive efforts were in the service of more broader normative ends. Therefore, today’s CE is actually a return to the roots of what analytic philosophy was destined to be from its start, before taking the infelicitous descriptivist turn in the 1960s and 1970s. To put it in Rortyan terms, Cappelen is offering a ‘redescription’²¹ of the history of analytic philosophy of language: analytic philosophy of language is not, as Soames writes in his influential depiction (Soames 2003a, 2003b), mostly a descriptivist enterprise, but it is and always was about the normative. The people I want to remain friends with would call that an attempt at counter-history, but to be clear, counter-histories are welcome and are what keeps progress going. In any case, Cappelen is enacting continuity by treating much of analytic philosophy’s historical figures as same-sayers with respect to his own enterprise.

For Rorty, the linguistic turn – the one he participated in the 1960s and whose roots go back to the beginning of the twentieth century – was an opportunity to shift philosophy’s focus away from the essentialist ambitions of metaphysics, and his works up to and including Philosophy and the Mirror of Nature (Rorty 1979) were declaredly an attempt to reform analytic philosophy from within (see Rorty 1992a; Rorty 1979, 7; see also Huetter-Almerigi and Ramberg 2022). Much like Cappelen, Rorty was drawing a line ‘from Broad to Smart, […] from Frege to Davidson, […] from Russell to Sellars, and […] from Carnap to Kuhn’ and insisted that this line ‘needs to be carried a few steps further’ (Rorty 1979, 7) – a project he undertook, in his opinion, with Philosophy and the Mirror of Nature. As stated above and as is well known, his work was not perceived that way. Ramberg and I believe that Rorty’s turn to more radical rhetorical strategies – pointing to (a certain degree of) rupture with analytic philosophy – was due to his growing sense that analytic philosophy in his day would not be hospitable to his views, which were mostly of a normative nature and bound to a picture of language where meaning is determined by both material and causal inferences and our current purposes; that is, what we need certain words for and what we want them to do for us (e.g. Rorty 1991, 5; Huetter-Almerigi 2020, 2022a).

The question for the future will be whether pointing to continuity will be enough for analytic philosophy to take in what I perceive to be the radical claims on Cappelen’s side, i.e. his conviction that there is no semantic or metasemantic touchstone that we will ever reach. I read Cappelen’s work as a constant attempt to have it both ways with respect to what Michael Losonsky in his Linguistic Turns in Modern Philosophy (Losonsky 2006)²² has described as the continuous tension in the history of philosophy, already present in Leibniz’s answer to Locke, between approaches to language that see language as

a systematic rule-governed structure or [as] […] an empirical object. If it is empirical, its domain and source of evidence are linguistic performance, but performance appears to elude the systematic determination. […] On the other hand, if language is seen as a system, its study will carve out a domain distinct from performance and empirical investigation. (Losonsky 2006, 251)

My reading is that Cappelen’s position in Fixing language (2018) serves the aim of accounting for both drives, leaving meaning what it is for large parts of mainstream analytic philosophy, i.e. a representational device defined by extensions and intensions which (traditionally, anyway) can be determined; while taking the performative aspect in through his notion of ‘topics’. Framed in a Quinean-Davidsonian vocabulary one could say that Cappelen moves indeterminacy to the level of topics while leaving meaning underdetermined. Therefore, while meaning, in theory, could be ‘found’,²³ the ‘finding-making’ distinction breaks down at least at the level of topics because, whether to ascribe continuity of topic or not has as much to do with ‘finding’ as with ‘making’.

I read much of Cappelen’s work along these lines; of trying to have it both ways, to maintain the connection with mainstream analytic philosophy and make a step beyond treasured certainties – e.g. his introduction of ‘content relativism’ instead of ‘truth-relativism’ (Cappelen 2008) or his recent defense of ordinary language philosophy (Cappelen and McKeever 2022) – to reform analytic philosophy from within. Future historians and philosophers will have to judge what came of it.

7. Conclusion

With respect to the topic of how to use our agency to bring about change for the better through linguistic interventions, which the historical traditions of American and Cambridge Pragmatism seem to share with CE, in this article I have pushed an openness in Cappelen’s position that allows for continuity with pragmatism. The reason is that the special issue that this article is part of aims for commensuration and establishment of common ground, whereas others might push for other lines. More generally I have argued that when setting out from a semantic and meta-semantic starting point like Cappelen’s, to go for or against topic-disruption is a question that needs to be answered by an appeal to what we want the chosen notion (‘rupture’ or ‘continuity’) to do for us.

My impetus was neither to argue for Cappelen’s insistence on continuity nor for Rorty’s insistence on rupture but to underline the possibility of choice between the two inside Cappelen’s (and Rorty’s) framework.²⁴ The upshot of my argument was that on purely theoretical grounds, Cappelen’s notion of ‘topics’ is giving us a convenient handle for ‘continuity in revision’ which is nothing less than the subject matter of large parts of conceptual engineering. However, once we have the notion of ‘topics’ in place, Cappelen’s insistence on topic-continuity is an optional addition which we can embrace or not. For Haslanger and Sterken, this is a choice inspired by considerations of political effectiveness. In the metaphilosophical context in which Cappelen and Rorty operate, it is a choice about how we want to describe philosophy, where what philosophy is proves to be an ethical and moral question: should philosophy as a discipline serve the aim of commensuration or help us endure and handle encounters with incommensurability, which might be more widespread than we might think or, in our desire for continuity, might hope for? Sometimes, continuity might be the right decision, and sometimes rupture might be the better option. In this sense, philosophy might well be at the same time a bunch of footnotes to Plato (Whitehead 1979) and progressing by changing the subject (Geuss 2017), hence contributing to the same conversation, but not the same inquiry.

Acknowledgements

Previous versions of this article greatly benefited from extensive comments by Céline Henne, Sigurd Jorem, Joey Pollock, Bjørn Ramberg, Matteo Santarelli and Matthew Shields, to all of whom I am immensely grateful. Obviously, none of the remaining shortcomings is theirs and they might not always be happy with what I made of their suggestions. A previous version of this paper was presented in September 2020 in the research network ‘Oslo Mind Group’. I am most grateful to the insightful comments received in that context. The final draft of this article was possible also due to a generous invitation from the University of Oslo’s Center for the Study of Mind in Nature research group in May 2022.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 Cappelen calls this the ‘Strawsonian Challenge’ referring to criticism moved by Strawson towards Carnap. I am not interested in whether Cappelen is construing the debate between Strawson and Carnap correctly (for critiques of this see Jorem 2021; Gascoigne 2022). I am interested in how Cappelen sets his argument up, what he makes of it, and what he needs it for.

2 There is no consensus in the literature about what CE is exactly operating on, whether it is concepts, word-meaning-pairs, representational devices, inferential devices, etc., nor what each of these aforementioned notions amount to. For various views see (Isaac 2020, 2021a, 2021b; Cappelen 2018; Haslanger 2020a). For a comprehensive list of positions see (Jorem and Löhr 2022) in this special issue, who also state their case for why we should view CE as engineering inferential devices. However, my topic here is Cappelen, and Cappelen does not believe in ‘concepts’ but only in ‘meaning’, where meaning is determined by extensions and intensions. On my reading, Cappelen introduces the category of ‘topics’, which is ‘compatible with changes in extension and intension’ (Cappelen 2018, 54), to circumvent the question of just how flexible meaning is. Put differently, someone’s ‘topics’ might as well be someone else’s ‘meaning’. I will come back to this point.

3 That is, the ‘Strawsonian Challenge’, see fn. 1.

4 That Rorty is not a ‘representational skeptic’ like Cappelen (2018, 6) but an anti-representationalist is not important for my argument at this point, though I do think the question of what the difference is between these two positions is an interesting one. On what anti-representationalism entails for Rorty see my (Huetter-Almerigi 2020, 2022a).

5 Or degeneration, for that matter; on non-inquiry or amelioration-driven conceptual engineering, see (Marques 2020; Podosky 2022; Shields 2021).

6 For a perspectival account of same-saying see Shields (2020), who then suggest that on his version of temporal internal/external perspectivism, one can actually do without the notion of ‘topics’ and still be able to account for change and/or continuity of meaning, respectively. I also take Shield’s account to be pragmatist in nature, as he thinks that what needs to be taken to overlap among speakers are purposes.

7 ‘Changing the subject’ in Haslanger’s sense might not be completely convergent with Cappelen’s ‘change of topic’. However, I believe that for the coarse-grained level of discourse that I am operating on the two are similar enough, and I think Cappelen could agree given that he himself cites Haslanger as one case of what he calls ‘topic-improving engineers’ (2018, 101 fn. 6, 148).

8 Haslanger has held different views over the years regarding amelioration vs. replacement. Further, given her Putnam-style semantic externalism, Haslanger also has theoretical (in addition to practical) reasons for insisting on continuity – on her account, in amelioration, we better grasp what the concept ‘really’ was all about all along, which seems to collapse amelioration into conceptual analysis (this has been noted before, by Cappelen 2018, 80; and Jackman 2020, 911). However, I am interested here in her practical (not theoretical) reasons.

9 Haslanger differentiates social from linguistic meaning (Haslanger 2012, 381–403).

10 On how this involves values and a value-concept circuit, see Santarelli (2022).

11 The inferential language here is mine; Haslanger refers to it as the sharing of ‘social meaning’.

12 On Rorty’s (congruent) take on the causal dimension of Davidson’s metaphors and the role they play in his account of language-renovation, see my (Huetter-Almerigi 2022b.)

13 If this unity then lies ‘only’ at the level of terminology, or if it encompasses content, topics, etc., is an interesting question that I owe to Joey Pollock. My intuition is that it is more on the mere level of terminology than we might like.

14 Another important venue for the study of lexical effects is literature. It is common among authors to have lists with words they want to use in their texts, where it is the very sound and rhythm of the word that make it interesting for use, not its semantic content.

15 That is, the choice of certain words and a certain terminology beyond or apart from their content, in as far as such a distinction can be usefully made.

16 Of course, there are many other differences, such as being an anti-representationalist vs. a representational skeptic for whom extension and intensions are equally malleable, among others.

17 Which is simply (nothing more than) our mutual attributions that we are still engaged in the same inquiry.

18 For Rorty and Fricker see Dieleman (2017) and Penelas (2019); for the general problem of speaking for others in Rorty see Voparil (2011b).

19 To fit Cappelen’s view, one would need to say ‘the commensuration of aims’. not meaning (if there really is a difference). As I will detail below, Cappelen introduces the category of ‘topics’ precisely to have it both ways: to allow for discontinuity of meaning (which is determined) and continuity of aims (which are performative). In Rorty, the two facets are integrated in his view of what language is and does.

20 For how incommensurability is not the same as untranslatability see Ramberg (1989). As Ramberg points out, conversation breakdowns are not the end of communication but the very starting point for radical interpretation. However, the end of radical interpretation might not lead to ‘unity of inquiry’.

21 To see what ‘redescription’ means for Rorty, and where the difference with respect to ‘redefinition’ lies, see (Ramberg 2006; Voparil 2011a). Roughly, ‘redescription’ is what we use to bring about changes in the world via linguistic means and where considerations regarding impact and purposes play a dominant role.

22 As with much of the rest of my analytic training I owe this reference to Bjørn Ramberg, without whom not.

23 Though one should further investigate both what is meant by ‘skeptic’ in Cappelen’s self-description as a ‘representational skeptic’ (2018, 6), and what that in turn implies with respect to the finding-making distinction.

24 For this openness between continuity and rupture in Rorty spelled out in terms of individual subjectivity, see my (Huetter-Almerigi 2022b.).