Conceptual engineering for analytic theology

ABSTRACT

Conceptual Engineering is the method (or methods) via which we can assess and improve our concepts. Can Conceptual Engineering be usefully employed within analytic theology? Given that analytic theology and analytic philosophy effectively share the same philosophical toolkit then if Conceptual Engineering works well in philosophy then it ought to work well in analytic theology too. This will be our working hypothesis. To make good on this hypothesis, we first address two challenges. The first challenge makes conceptual engineering look to be too inclusive; the second challenge makes it look to be too revolutionary (for analytic theology). To address these challenges, we propose a refined characterisation of Conceptual Engineering. We then turn to consider a number of case studies where analytic theology and conceptual engineering may fruitfully cooperate. These are: theological disagreements, inter-faith dialogue, meaning change, celibacy, AI, the name of God and conceptual genealogy.

KEYWORDS:

• Conceptual engineering
• analytic theology
• conceptual pluralism
• metalinguistic negotiation
• religious disagreement
• inter-faith dialogue

1. Our working hypothesis

Conceptual Engineering is the method (or methods) via which we can assess and improve our concepts.¹ Given some conceptual problem, you make progress by suitably revising (or replacing) the concept(s) which give rise to the issue. Could this method be applied within philosophical theology? Could we use this method to address questions concerning the nature and existence of God? Likewise, for debates concerning the soul, the nature of evil, religious experience and more?

Our focus is on a more specific query: Can Conceptual Engineering be usefully employed within analytic theology? This question is more tractable because although Conceptual Engineering is a general method for addressing conceptual problems, it has risen to prominence within analytic philosophy. Given that analytic theology and analytic philosophy effectively share the same philosophical toolkit, then if Conceptual Engineering works well in philosophy, it ought to work well in analytic theology too. This will be our working hypothesis.

2. Two immediate challenges

This hypothesis, however, faces two challenges. The first emerges from the observation that not only is Conceptual Engineering a new movement (with seemingly little tradition) it also seems to promote a highly revisionary method. In caricature form: out with the old concepts and in with the new. This conception of Conceptual Engineering is prefigured in Nietzsche’s remarks that philosophers.

[…] have trusted in concepts as completely as they have mistrusted the senses: they have not stopped to consider that concepts and words are our inheritance from ages in which thinking was very modest and unclear. … What dawns on philosophers last of all: they must no longer accept concepts as a gift, nor merely purify and polish them, but first make and create them, present them and make them convincing. Hitherto one has generally trusted one's concepts as if they were a wonderful dowry from some sort of wonderland: but they are, after all, the inheritance from our most remote, most foolish as well as most intelligent ancestors. … What is needed above all is an absolute skepticism toward all inherited concepts. (Emphasis original. 1968, 220–221)²

If Conceptual Engineering is about moving beyond (or being absolutely sceptical about) the concepts of old then it is anathema to analytic theology. After all, analytic theology betrays no overt interest in moving beyond the central theological concepts such as the concept of God, the concept of the afterlife, the concept of faith and kindred concepts. Typically, theology seeks to retain these concepts and instead gain a deeper understanding of them.³ Call this The Revolutionary Challenge, the challenge that Conceptual Engineering is too revolutionary for analytic theology.⁴

A second challenge is that the Conceptual Engineering movement tends to portray every philosopher as a Conceptual Engineer. Upshot: the Conceptual Engineering movement turns out to be an empty exercise in re-branding. If this were so, then our main question becomes: Can (analytic) philosophy be usefully employed within analytic theology? The answer is, of course, a trivial: Yes. Call this: The Re-Branding Challenge. To be of interest, Conceptual Engineering had better be offering up new (or at least neglected) ways of doing philosophy such that not every philosopher turns out to be a Conceptual Engineer.

3. Goals

With respect to the Revolutionary Challenge, in §6, we provide various responses which leave plenty of scope for analytic theology to deploy the tools of Conceptual Engineering. The lesson is that Conceptual Engineering yields a more diverse and modest methodology than the challenge suggests – one which does not enforce the edict that we should do away with (or be absolutely sceptical about) our inherited concepts. With respect to the Re-Branding Challenge, in §§8–9, we propose a revised characterisation of Conceptual Engineering under which not every philosopher is a Conceptual Engineer.

In §§10–16 we then survey seven case studies where Conceptual Engineering and analytic theology may fruitfully cooperate. These are: theological disagreement; meaning change for religious terms; inter-faith dialogue; the name of God; conceptual genealogy; AI and theology; and celibacy. Our purpose is not to provide definitive answers, but rather to open up promising avenues of cooperation between analytic theology and conceptual engineering.

4. What is conceptual engineering?

On a broad conception, Conceptual Engineering is the method (or methods) via which we assess and improve our representational devices.⁵ This formulation allows for different kinds of Conceptual Engineering since the primary representational devices could be: concepts, words, meanings, or the contents of thoughts. In turn, this plurality is further multiplied given the various conceptions of the metaphysics of words, concepts, meanings, contents.⁶ More kinds of Conceptual Engineering emerge given the different sorts of philosophical questions/problems to be addressed.⁷ Likewise for the various reasons for improving some representation device.⁸ We can’t survey (or evaluate) all the options here. Instead, we shall outline five promising kinds of Conceptual Engineering found in contemporary literature.

4.1. Replacement conceptual engineering

The primary representational devices are concepts (and concept-words). You make headway on philosophical problems by replacing the old concept-word at large with a new concept-word (which expresses a new, better concept).⁹

Case Study One: The puzzles surrounding the nature of knowledge show that the concept of knowledge is defective – typically that the concept is too demanding such that knowledge is impossible. We should replace this concept with some suitable (less-demanding) surrogate and use a different word to pick this new concept out. (See (Schiffer 2004) on replacing ‘justification’ with ‘justification*’; cf. (Sider 2009) on ‘Ontologese’.)

Case Study Two: The semantic paradoxes show that the concept of truth is defective (indeed ‘incoherent’). This concept should be replaced with a less defective (‘coherent’) surrogate which goes by a new name (See Tarski (1944, 356) on ‘true’ versus ‘frue’; Scharp (2007, 2013) proposes that the concept needs replacing by two surrogates.)

4.2. Revisionary semantic engineering

On a more moderate, revisionary form of Conceptual Engineering, a new word is not needed. Instead, the old word gets a better meaning:

Revisionary Semantic Engineering: The primary representational devices are words (and meanings).¹⁰ You make headway on some philosophical issue by giving a word such as ‘belief’, ‘consciousness’, ‘knowledge’ a new and better meaning. To do this, the use of the old word must be suitably changed. (Cappelen 2018 is the locus classicus of this view.) See also (Cappelen 2018; Cappelen and Plunkett 2020; Thomasson n.d.; Pinder 2021) (Cf. Carnap 1950, chap. 1).

Case Study Three: Take the debates concerning the nature of belief. The old meaning of ‘belief’ is holding these debates back. Either because it is not picking out a sufficiently unified kind or because it is not picking out the right kind. For example, it can be argued that that it is part the old meaning that beliefs are ‘in the head’. However, there are reasons to think that beliefs may, in part, be determined by external factors. So, we need to give the word ‘belief’ a new and better meaning to capture this feature, otherwise the word will not properly latch onto all and only beliefs. We do this by changing the use of ‘belief’. (See (Cappelen 2018, 10) who interprets Clark and Chalmers (1998) as engaging in Semantic Engineering.)

4.3. Conceptual pluralism

The primary representational devices are concepts (and concept words).¹¹ A typical philosophical conflict exhibits a kind of verbal disagreement whereby the locus of disagreement turns on some meta-linguistic matter (such as what a word used in the dispute means). To resolve the issue, you must attend to the multiple, related, over-lapping conceptual roles (or meanings) that are attached to the original concept.

Progress on the problem takes place by adopting ‘the subscript-strategy’ (or some functional equivalent) whereby the concept-term ‘F’ is replaced with a group of terms ‘F₁’, ‘F₂’, ‘F₃’ … , which are each assigned one of the conceptual roles (meanings) which are attached to the original concept. This allows the original dispute to dissolve because we should no longer use the term ‘F’ but instead deploy the surrogate terms which each have reasonably clear conceptual roles (meanings). (See Chalmers (2011).)¹²

Case Study Four: Take the term ‘meaning’. There is a long-standing debate over the nature of meaning. One theorist says: the meaning of a term is its reference (namely, extension). Another theorist says: meaning is that thing which fixes reference (namely intension). A further theorist says two words can have the same intension and yet differ in meaning (because they differ in ‘cognitive value’). In this case, there is no need to use the subscript strategy because we have three replacing terms to hand: ‘extension’, ‘intension’, ‘cognitive value’. The thought goes that we no longer need to debate the meaning of ‘meaning’, but rather deploy the replacing concept-words/concepts instead to play the separate conceptual roles that were attached to the original concept.

4.4. Metalinguistic negotiation

The primary representational devices are concepts (and concept-words).¹³ A philosophical debate about the nature of knowledge, belief, consciousness and so on, turns out to be (in part) a meta-linguistic debate concerning what meaning/concept should be attached to some word. Making this normative, meta-linguistic dimension salient to the disputants puts them in a (better) position to resolve matters by suitably negotiating how the term should be used. The disputants may well be able to agree on a series of conditionals of the form: if your goal is to have a concept which plays role R₁ then it should have meaning M₁; if your goal is to have a concept that plays role R₂ then it should have meaning M₂. (See Belleri 2020; Mankowitz 2021; Plunkett 2015; Plunkett and Sundell 2013; 2021; Thomasson 2017.)

Case Study Five: A case study in Plunkett (2015) involves two speakers who disagree as to whether there is free will. The locus of this dispute is meta-linguistic (despite initially seeming otherwise): it is, in part, a disagreement about what the term ‘free will’ should mean. The dispute then becomes more tractable because there may be some common ground at large in the form of the conditionals: if the concept of free will is to mesh with our everyday attributions of responsibility then the meaning of ‘free will’ should fix a low-bar for attributions of free will (such that free will is not threatened by determinism or indeterminism). If the concept of free will does not need to respect everyday attributions of responsibility then the meaning of ‘free will’ should fix a high-bar for what counts as free will. This common ground then allows for a more productive debate as to what we want the concept to be.

4.5. Conceptual/semantic design

The four forms of Conceptual Engineering given above agree that making progress on philosophical questions involves implementing the requisite changes in revision/replacement. This gives rise to the ‘Implementation Challenge’. When directed at revisionary forms of Conceptual Engineering, this is: the factors which determine which meaning or concept is expressed by some concept-word are either not easy to control; or, we don’t have a complete picture as to how the meaning/concept expressed by a word is determined by use; (or both).¹⁴

Cappelen (2018, 74 ff.) makes a plausible case that we can be cautiously optimistic about implementing the requisite conceptual/linguistic/semantic changes. That’s because, like many purposive activities (such as child-rearing), we don’t need to possess a detailed instruction manual in order to succeed. Rather, we just try to bring about the requisite changes (typically by employing rules of thumb) and indeed we often do succeed. While this reply is promising, and is certainly needed for applied forms of Conceptual Engineering which aim to bring about some practical goal, it is arguably not needed to allow for progress in philosophy. How so?

Because the Challenge over-states the success-conditions for a project in philosophy. In particular: implementing the proposed revisions or replacements is not needed. Rather, all that is required is that one or more promising blueprints for conceptual or semantic change is provided. In other words, a promising conceptual/semantic design (delivered in the meta-language) is what is required. Conceptual Engineering (in philosophy) is better deemed to be: Conceptual Design.¹⁵

Case Study Seven: To illustrate: suppose you wish to address some philosophical conflict over moral responsibility. One blueprint for semantic change recommends that the meaning of ‘responsible’ be such that some range of attributions of moral responsibility (expressed in the meta language) come out as true. Another blueprint recommends a different meaning for ‘responsible’ under which not all of these attributions come out as true. These two designs can then be stress-tested in various ways (e.g. by thought experiments) in order to establish which offers the most promising blueprint for semantic change. These stress-tests, if they are comprehensive enough, will indicate that one design is better than the other. So, we can make progress on a philosophical problem without implementing the proposed changes.

5. What is analytic theology?

The most bare-bones articulation of analytic theology is: the study of theology that makes use of the methods, tools and concepts generally employed in analytic philosophy. Following Wood (2021b, 50), one way to construe analytic theology is merely as another instance of theologians making use of whatever is the prevalent and popular philosophical tradition at the time. While the ‘official’ canon of analytic theology begins with Crisp and Rea’s (2009) edited volume, there is agreement that analytic theology more broadly interpreted had begun decades prior, in the work of ‘philosophers like Alvin Plantinga, Richard Swinburne, Robert Adams, Marilyn McCord Adams, Nicholas Wolterstorff, Eleonore Stump and others who, in the latter half of the twentieth century, played a major role in the revival of philosophy of religion and the growth of philosophical theology within academic philosophy’ (Rea 2022).¹⁶ Hence, if analytic philosophy is sometimes characterised as being born of a ‘linguistic turn,’ we can equally speak of another ‘turn’ within analytic philosophy towards philosophy of religion and theology (Macdonald 2014, 33–34).

One finds a variety of ways in which analytic theology is understood and put into practice. Firstly, there is debate as to where (or whether) one should carve out the delineation between analytic theology, and systematic theology on the one hand, and philosophy of religion and theological philosophy on the other.¹⁷ This is precisely because we see analytic theology exploring a wide variety of questions, from more philosophical ones such as the existence and essence of God, to issues of a more doctrinal and dogmatic nature concerning the Trinity, atonement and the person of Jesus Christ. Beyond this, there are also mixed views as to whether there are some fixed theological commitments that are essential to analytic theology. Analytic theology should certainly not be considered as some monolithic project having a single goal. While it might be the case that a larger part of the literature and work being carried out within this field is within the Christian tradition, we also see analytic theology being applied and taken up within other religious traditions such as in Jewish¹⁸ and Islamic theology.¹⁹

What suffices for our goals is that analytic theology lies on the overlap and interface of philosophy and theology. There is, however, another crucial point to note. It would be too reductive to simply assert that analytic theology has merely adopted the style employed by analytic philosophy. Analytic theology isn’t called as such because it ‘looks like’ analytic philosophy. The link between the two is deeper; works in analytic theology also make use of concepts, theories, presuppositions and frameworks employed in analytic philosophy.²⁰ This is reminiscent of the old dictum of philosophia ancilla theologiae.

All this already goes some way in justifying the main question we have set out to explore in this paper. We can now proceed to show how analytic theology also engages and grapples with different kinds of questions; from those of a more theoretical nature, to more practical and social issues.²¹

6. The revolutionary challenge

The Revolutionary Challenge is the challenge that Conceptual Engineering is too revisionary to be useful to analytic theology. One response is to insist that analytic theology should also become a highly revisionary enterprise. While there are theological traditions marked by revolutionary ideas, this kind of response is misplaced.²² That’s because Conceptual Engineering, when properly conceived, isn’t unduly revolutionary – or so we shall argue.

Futurism vs Revivalism: Futuristic proponents of Conceptual Engineering tell us that it is a new and exciting way of doing philosophy (with little in the way of tradition).²³ For these advocates, philosophy in the past has been too descriptive. If philosophical problems are to be resolved then we need a prescriptive turn. If such Futurism was an inherent feature of Conceptual Engineering then the Revolutionary Challenge would indeed get a grip. On a revivalist view, in contrast, what is needed, is a prescriptive return in philosophy. Conceptual Engineering is not something that is particularly new. It has had practitioners in the past – though at times it was being done inadvertently. Recently, it has fallen out of fashion in favour of descriptive forms of philosophy. Conceptual Engineering, on this view, offers a template for bringing back these past, neglected forms of prescriptive philosophy.²⁴ Conceptual Engineering thus need not be seen as yielding a new way of doing philosophy.²⁵

Global vs Local: Even if you did think that Conceptual Engineering offers a radical method for doing philosophy, you might nonetheless sponsor a local version of Conceptual Engineering whereby it is only applied to certain forms of philosophy or philosophical problem.²⁶ (Contrast: a global version where it is going on all the time.) So, for example, you might think that it is only to be applied as a kind of last-resort tool to various intractable puzzles.²⁷ An analytic theologian could readily adopt a localised but futuristic form of Conceptual Engineering to address some theological paradox, without thereby advocating for a radical vision for theology.

Assessment is Descriptive: It’s worth noting that global forms of Conceptual Engineering (whereby all philosophical activity involves revision/replacement) are in any case suspect. That’s because there are two basic stages to the engineering process: description (What is the nature of this meaning/concept/concept-word?) and then prescription (What should the concept/meaning be?). So, there is always a purely descriptive element to the process. Conceptual Engineering is thus less revolutionary than one might initially imagine because the revision (or replacement) is not taking place during assessment.

Prescriptive Neutrality: Most characterisations of Conceptual Engineering (including that given in §1) tend to suggest that philosophical progress is made by the revision or replacement of our representational devices. This is somewhat misleading. The doctrine should allow that the assessment could yield the verdict: no revision or replacement is needed. Perhaps the problem being addressed is illusory; perhaps the issue can be addressed by other means; or perhaps the proposed revisions are too costly.²⁸ This is a familiar predicament: there can be progress on a problem by discovering that preserving the status quo is the best response. More generally, Conceptual Engineering needs to be prescriptively neutral such that it is an open question (in advance) as to whether revision/replacement is needed. This neutrality makes for a much less radical doctrine.

Methodological Scepticism: Nietzsche, as we saw in §2, calls for an absolute scepticism as regards our inherited concepts. There is some merit to this edict because the suggested stance is not the revolutionary stance of recommending these concepts be abandoned but is rather (on a plausible reading) the stance of open-mindedness as to whether our representational devices are doing well. Analytic philosophers and theologians need to fully stress-test our inherited concepts to see whether they require revision or replacement. Read this way, Nietzsche is merely recommending a methodological scepticism akin to the ‘clean sweep’ proposed by Descartes in the Meditations. There is nothing inherently revolutionary about that – indeed embeds the prescriptive neutrality just mentioned.

Revision rather Replacement: Of the five kinds of Conceptual Engineering outlined in §4, Replacement Conceptual Engineering is the most radical. It recommends both replacement of concepts (or meanings) and replacement of concept-words. Sometimes this feature can be obscured. For example, Scharp (2013) proposes that we replace the concept of truth, and the word ‘true’, with the concepts of ascending truth and descending truth (together with the corresponding terms). This terminology suggests that these replacement concepts are types of truth. They are not! Scharp is not deploying a hitherto unnoticed distinction between two kinds of truth.²⁹ So, his view does not admit the definition: x is true if and only if x is ascending true or descending true. If it did, we could keep the concept of truth. This is why his proposal (and kindred proposals) is radical: it constitutes a kind of Philosophical NewSpeak whereby the concepts that we learnt at our mother’s knee (‘truth’, ‘knowledge’, ‘goodness’) are to be replaced if we wish to make progress.³⁰

Likewise, when Schiffer (2004) proposes replacing the concept of justification with the (less demanding) concept of justification*, don't think of this latter concept as denoting a kind of justification. It denotes a surrogate for justification. Underpinning these views, is a fine-grained view of concepts (whereby a small difference between concept A and concept B entails these concepts are distinct) together with some naming principle of the form: a new concept needs a new name. These principles are, arguably, highly problematic.³¹ The analytic theologian need not embrace these radical forms of the replacement view.

Instead, what is arguably needed is a coarse-grained view of concepts which allows for both revision and replacement. On this view, a substantial revision of a concept produces a new concept (whence a new name is needed), while a minor revision allows the concept to survive the process of alteration. This ecumenical view allows some problems may require substantial revisions, while others merely need less substantial conceptual changes.³²

Surrogacy: In typical cases of conceptual revision/replacement, the revised or replacing concept serves as a surrogate for the old (version of the) concept. In slogan form: do the job of the original concept but do it better. This entails that most features of the revised/replacing concept will be preserved. This feature means that typical cases of Conceptual Engineering are, again, not as disruptive as you might think.

Monism vs Pluralism: We have just seen how some forms of Replacement Conceptual Engineering sponsor a fine-grained view of concepts (Scharp 2013, 2019, 2021). On these views, it’s replacement or nothing! That represents a monistic conception of Conceptual Engineering. In contrast, the more ecumenical view mooted above permits both conceptual replacement and revision. That is more pluralist. Since the jury is still out as to which form of Conceptual Engineering is the most promising, we recommend methodological pluralism whereby each of the versions of Conceptual Engineering sketched in §4 is in the running. After all, it’s a standard feature of engineering methodology to propose more than one kind of solution to some problem. With such pluralism in hand, the Revolutionary Challenge is even less pressing because the analytic theologian is not committed to deploy the (somewhat) more radical versions of Conceptual Engineering.

Conceptual Design: Finally, it was floated above that Conceptual Design provides the best way to address the Implementation Challenge (for philosophical projects). Once the implementation stage is not seen as essential then this also serves to lessen the Revolutionary Challenge. That’s because one reason why Conceptual Engineering can seem to be so disruptive is because it is typically thought that we have to bring about the requisite revisions/replacements to make progress. If Conceptual Engineering is all about producing (in the meta-language) one or more suitable blueprints for revision or replacement then it is far less disruptive.

The considerations just given collectively suggest that Conceptual Engineering, despite being a new movement, really isn’t offering up a revolutionary methodology. Our working hypothesis thus remains a live option.

7. A broader bridge?

The goal of the previous section was to clear the way for a bridge between analytic philosophy and analytic theology – by way of conceptual Engineering. This is not the only bridge between these disciplines. There is growing cooperation with respect to metaphysics, epistemology, and, more recently, on theological paradox.³³ Given these broader kinds of connection, one concern is that the bridge being proposed is too narrow – why not simply highlight the under-explored route which goes by the way of the philosophy of language? Since Conceptual Engineering is, largely, embedded in debates in the philosophy of language (over the nature of words, the nature of meaning, meaning change, meta-semantics, disagreement, the nature of concepts and so on), then this broader bridge subsumes the narrower route being proposed.

Our answer is: by all means, let’s explore this broader bridge too. Indeed, it is somewhat surprising to find that while contemporary analytic theology has reconnected to debates within metaphysics and epistemology, it has yet to engage with debates that have taken place within the philosophy of language over the last 50 years. While this is understandable, given the dominance of Logical Positivism in the first half of the twentieth Century, pre-theoretically one would expect people working within Christian religious philosophy to look toward philosophy of language with tremendous interest. After all, Christianity, is a religion is based on the Word.

The need for such a broader bridge does not, however, make the proposal of this paper redundant. Going by way of Conceptual Engineering is just one more way in which analytic theology and philosophy may fruitfully cooperate. Furthermore, even those who might be sceptical of the Conceptual Engineering movement as a meta-philosophical project can nonetheless benefit from seeing how the many new insights and ideas that Conceptual Engineers have brought to the philosophical table may help illuminate issues within analytic theology. Before turning to these areas of cooperation, we must first address the Re-Branding Challenge.

8. The re-branding challenge

This is the challenge that the Conceptual Engineering movement is just an empty exercise in rebranding philosophy. If so, it’s just trivial that that Conceptual Engineering can help analytic theology since everybody can agree that analytic philosophy is of service to analytic theology. This Challenge emerges because while Conceptual Engineers are typically very good at telling us what Conceptual Engineering is, by way of giving us typical exemplars, they are not always so good at telling us what it is not. This is the extensional version of the challenge: too many kinds of philosophy get to be included in the extension. There is also an intensional version of the challenge: the most serviceable definition of ‘Conceptual Engineering’, as the method via which we assess and improve our representational devices, is far too inclusive – because it ends up including all forms of philosophy. We shall merely be concerned with this intensional version.³⁴

Beliefs are essentially states that represent how things stand in the world. As such, beliefs are essentially representational devices. Given the leading definition of ‘Conceptual Engineering’, which, again, says that it is the method via which we assess and improve our representational devices, then improving our beliefs counts as a form of Conceptual Engineering. Belief improvement may take many different forms: improving the content of a belief; improving the mechanism by which a belief was formed; improving its evidential grounding; improving its rational status; improving its strength; and more. More generally, sets of beliefs are representational devices too, and we can improve this set, for example, by getting rid of some belief. Such Belief Engineering is very familiar. Every philosopher is a Belief Engineer, and obviously so. You believe that the concept of truth is inconsistent, for example, while I try to (philosophically) persuade you not to have this belief. The upshot is that the most promising characterisation of Conceptual Engineering is too broad because it allows Belief Engineering to be a form of Conceptual Engineering – and so lands us with the Re-Branding challenge.

Where does this leave us? Well, Conceptual Engineers are certainly on the right track when they say that Conceptual Engineering is about assessing and improving our representational devices. A revised characterisation is nearby.

9. Conceptual engineering re-characterised

We can now kill five birds with one stone. We can take into account the feature of prescriptive neutrality outlined in §6; we can be open-minded about which representational devices are the primary objects of Conceptual Engineering (methodological pluralism); we can be open-minded about whether revision or replacement is called for (more methodological pluralism); we can take into account that Conceptual Engineering is best conceived as Conceptual (or Semantic) Design; and, finally, we can take into account that mere Belief Engineering does not entail Conceptual Engineering. This yields the following characterisation:

Working Definition: Conceptual Engineering consists of (1) the methods via which we assess our primary representational devices to see whether they are in need of re-design, together with (2) the methods via which a design (for revising or replacing these devices) can be produced and evaluated as acceptable.

Here the primary representational devices are: words, meanings, concepts and the contents of propositional attitudes. There are of course derivative, secondary representational devices such as beliefs and other propositional attitudes. These are representational in a derivative sense because it is only via having a content that they get to be representational. These are excluded from the scope of Conceptual Engineering – thus avoiding the Re-Branding Challenge. Furthermore, the re-design might involve producing blueprints for change that involve revision or replacement (or Conceptual Pluralism or Metalinguistic Negotiation).

It is this working definition of Conceptual Engineering that should be plugged into our working hypothesis: if Conceptual Engineering works well in philosophy then it ought to work well in analytic theology.

10. Theological disagreement

Consider a Christian theological dispute about disability. On the one hand, there are Christians who endorse the literal deliverances of the Bible and so take God to have created deafness, blindness and indeed all human disabilities.³⁵ On the other, there are Christians who allege that disabilities arise because of the Fall (see Genesis 3). This latter group appeal, amongst other things, to argument: how can imperfection (disability) have come from perfection (God)? One way to address this dispute is to turn to the arguments concerning God’s creativity and perfection. Another way is via the epistemology of scripture: is scripture infallible? Conceptual Engineers, meanwhile, offer a different approach.

Take Conceptual Pluralism. On this view, if we have good reason to suspect that a debate is intractable then we should investigate whether the locus of the dispute is merely verbal – whether it grounds out in some dispute, or misunderstanding, concerning what one or more of the terms in the dispute means. The terms under such dispute, or misunderstanding, might be: ‘God’, ‘the Fall’, ‘the Bible’, ‘the word of God’, ‘scripture’, ‘perfection’, ‘creativity’, ‘infallibility’, and also ‘blindness’, ‘deafness’, ‘disability’. The Conceptual Pluralist begins by being open-minded about which of these terms is the (potential) locus of the verbal dispute. Nonetheless, for our purposes, it will be instructive to select the term ‘disability’ for scrutiny. Is there some conflict and/or misunderstanding about what this term means?

The disputants may well be able to reliably categorise disabled people: the blind, the deaf, those with a learning disability, those with an orthopaedic disability, and so on. Further scrutiny, however, shows that they associate the concept of disability with multiple, over-lapping and conflicting concepts: the concept of impairment, the concept of defect, the concept of low function, the concept of abnormality, the concept of non well-being, the concept of maladaptation. Even further scrutiny, reveals that these concepts differ with respect to two key (related) issues: Is disability intrinsically harmful? Does disability essentially lead to a less flourishing life?

Recent (philosophical) work on disability casts doubt on the common view that disability is intrinsically harmful, that disability automatically yields a less flourishing life.³⁶ Being maladapted does not necessarily produce a less flourishing life; being impaired might; and having a defective body (or mind) typically does. So, the multiple over-lapping concepts we associate with disability are in conflict. The Conceptual Pluralist suggests we should adopt a strategy of divide and conquer: in addition to the concepts listed above, we should also introduce one or more new replacement concepts. One of these is: the concept of soma-divergency. This concept, together with the more familiar concept of neuro-divergency, can stand as a suitable replacement for the concept of being disabled. Being soma-divergent just means, to borrow Barnes’ terminology, you have a minority body (roughly, a type of body that not every person has); and being neuro-divergent just means, to extend Barnes’ terminology, you have a minority mind (roughly, a type of mind that not every person has).

How does this help with our puzzle? The Conceptual Pluralist, firstly, notes that both parties to the dispute agree that disability is a less flourishing state. Secondly, they note that the concept of disability (on our current messy understanding) does not deliver a clear verdict that disability is intrinsically harmful or automatically leads to a non-flourishing state. If this concept gets unpacked via the concept of defect, the verdict would be: yes. If the concept of disability gets unpacked via mere maladaptation or abnormality then the answer is: not necessarily. So, those who believe that disability is an intrinsically bad thing which must have been produced by some malign force need to refrain from making these very claims. That is not a loss for one side, and a win for the other, however. While the Biblical Literalists can retain their belief that God created deafness, blindness and so on, they cannot go on to claim that God created disability. Again, that’s because the concept of disability is messy: there is no clear verdict that disability is intrinsically harmful. This represents a kind of progress on the dispute because we can diagnose that both sides are at fault.³⁷

However, the Conceptual Pluralist then suggests that both sides can usefully deploy the new concepts of neuro-divergency and soma-divergency. Both sides should agree that God created neuro-divergent people and soma-divergent people and that he created blind, the deaf, those with Asperger's syndrome, and so on. That might, however, seem like a mild win for the Biblical Literalists because they get to retain their claim that God created deaf people while their opponents must give up their claim that God did not create deaf people. It’s not an overall win, however, because there are people which we can properly identify as thoroughly prevented from living any kind of flourishing life (because all their sensory modalities are absolutely impaired). The Bible does not say that these people are created by God. Rather, it remains a live option to argue that these people are the way they because of our fallen state. So, it turns out that the Bible Literalists must give up their claim that God has created the thoroughly impaired. In fact, other things being equal, this will turn out to be common ground between both parties. So, there is compromise on both sides.

Conceptual Pluralism is mildly radical. The requisite conceptual and linguistic changes might be deemed too disruptive. You could instead seek to address this dispute by using Semantic Engineering whereby the goal is to revise the meaning of ‘disabled’ such that the following definition comes out as true: ‘x is disabled  = _df x is soma-divergent or neuro-divergent’.³⁸ On this proposal, the word ‘disability’, as currently used, has a bad meaning: it excludes people who have a flourishing life. That’s because on its current meaning the sentence ‘Disability is intrinsically bad’ comes out as true. On its improved meaning, however, this sentence comes out as false – thus allowing us to not to treat disabled people and those who have a flourishing life as mutually exclusive groups. How does this proposal help?

It suggests that those believers who oppose the Biblical literalists may well be right – given the current meaning of ‘disabled’. Their claim that God did not create disabilities is true (for all parties), given the current meaning of ‘disabled’. However, there is a sense in which this should not be the meaning of the term. Given its improved meaning, the sentence ‘God did not create disabilities’ comes out as false. The upshot, then, is just a temporary win for the non-literalist Christians. It’s an overall loss because they only get to be able to say something true here by using the wrong meaning of ‘disabled’. Likewise, it's an overall win for the Biblical Literalists because, while they do say something false, they only do so because they are forced to use the current, bad meaning of ‘disabled’. Were they to be able to use the improved meaning, they would be able to say something true.

11. Celibacy

Above we have seen cases where concepts were problematic due to long-standing paradoxes involving truth or justification. The case of celibacy, however, provides us with rather different class of challenge: the shift in meaning is happening ‘right before our eyes’ as it were. What does the Conceptual Engineer have to say about this kind of case?

Celibacy is a long-standing religious practice found in many different religious traditions.³⁹ From Catholic priests, to Buddhist monks and nuns, to Hindu sadhus, celibacy is part of a wider way of life that such individuals choose to adopt and live by. A bare-bones way of articulating celibacy is as a specific type of vow or promise made to not engage in sexual activity for some religious purpose. There are some important points that one should underline in this definition. Whilst appreciating the differences that may be present in the way specific traditions practice and understand celibacy, some aspects seem to be universally shared. For example, given the practice’s insertion in a wider religious (and, hence, moral) framework, celibacy is generally understood as not only consisting in the choice to forgo having sexual partners, or engaging in physical sexual acts, but also in seeking to guide one’s thoughts and desires away from the pleasures derived from such activity. Furthermore, celibacy – inasmuch as it is usually linked to a specific role or position within a religious tradition – is generally taken up indefinitely. In this regard it can be likened to marriage, in that marriage is not usually undertaken for a pre-specified time-bound period.

At the same time, however, the concept of celibacy being employed within non-theological, contemporary contexts is very different – and conflicting. Recent years have seen the emergence of individuals who identify as ‘incels’ – involuntary celibates.⁴⁰ Continually frustrated in their endeavours to find a sexual partner, men who identify as incels subscribe to a particular narrative as to why they find themselves in this situation (Preston, Halpin, and Maguire 2021), and believe that violence is the only solution. Analogously, there is the less popular term of ‘femcels’ (though this group does not exhibit the level of violence and hatred characteristic of incels). Furthermore, other novel terms that are being occasionally used are ‘unwanted celibacy’ (Grunau et al. 2022), ‘voluntary celibacy’ (Saner 2023) and ‘accidental celibacy’ (Brooks 2019; Greenstreet 2021).

The concept of celibacy as employed in the above examples seem to differ from the traditional meaning of ‘celibacy’ in a number of significant ways; there does not seem to be a link between celibacy and a higher, religious purpose; it is understood simply as not engaging in physical sexual activity with another individual, and – more importantly – it seems that it is not implicit within the very definition of celibacy that this is a voluntary choice. On this final point, it is plausible to assume that theologians would be unwilling to characterise the above phenomena being described as actually counting as instances of celibacy.

We therefore have a conflict and Conceptual Engineering might well have the tools to resolve it. One may choose the route of the Conceptual Pluralist and carve out two or more different notions of celibacy, celibacy₁ which captures the traditional, religious understanding of the concept, and celibacy₂ which refers to the contemporary usage. An alternative step would be to ensure that one or the other use of the concept should be the only meaning; choosing either to stick to the traditional understanding and maintain the status quo, or instead to revise the concept and adopt the new modern usage. Or perhaps the original concept has been so degraded that complete conceptual replacement is called for. Whichever route is chosen depends to a large degree on the goals the particular theologian-cum-conceptual engineer has in mind.

12. The semantic commandment paradox

Let's now consider a theological paradox involving meaning change. This time, however, our goal is not to exhibit how various Conceptual Engineers might deploy their favoured template for resolving a problem. We hope the previous two sections have done enough to convince you that these Engineers are bringing something new and worthy to the meta-philosophical and meta-theological table. Rather, our goal is to reveal how Conceptual Engineers also bring illuminating ideas and concepts to the first-order philosophical table.

At several places in the Bible, we are commanded not to add or subtract words from the text.⁴¹ That command precludes replacing a biblical word with a surrogate word with a very similar meaning. It arguably also commands us to preserve the meaning of a biblical word, since using some biblical word in a way which does not line up with the meaning the word had when the book was written diminishes the word. The purpose of these commandments is easy to appreciate – there is but one inviolable text, with one fixed meaning.⁴² This raises a puzzle. Any piece of scripture is written in natural language. The meaning of the words in natural language can and do change. Indeed, it is effectively impossible to prevent the meaning of a word in natural language from changing. In effect, the Bible commands us to do the impossible: do not change the meaning of biblical words. Paradox! What do Conceptual Engineers have to say?

Semantic Engineers tell us that the phenomenon of meaning change is already paradoxical for independent reasons.⁴³ To see why, note that in correctly reporting what you said (at some other time or context) I can simply re-use the sentence you uttered in reporting the content of your speech. To illustrate, suppose my ancestor uttered the sentence ‘God is merciful’ in 1500. Suppose I know this and report what my ancestor said as follows: ‘My ancestor said that God is merciful’. Suppose I now utter the sentence ‘God is merciful’ and so I can now correctly offer the following same-saying reports: ‘My ancestor and I said the same thing’, ‘What we said was the same.’ If we said the same thing, then our words have the same meaning across the interval of time. Upshot: no meaning change!

This is puzzling since the meaning of ‘merciful’ has surely undergone some change in meaning over this interval. This word seems to have originally just meant ‘refraining from delivering a deserved punishment’; now it has acquired a more involved meaning whereby it comes about because of kindness, compassion or forgiveness. More generally, it’s arguable that just as this word could have had a slightly different meaning when this word first took on its meaning in English, there is a range of easily possible meanings at each moment of its usage which it could easily latch onto. After all, the word gets to be used slightly differently from one moment to the next, so this usage will tend to issue in a change in meaning over time. So, the meaning change of ‘merciful’ is not just natural, it is effectively inevitable. This is paradoxical because the speech reports given above seem to enforce that there is no difference in what is said, and therefore what is meant, over time by a use of the sentence ‘God is merciful’.

Here the Semantic Engineer has an important insight which emerges from their answer to another problem – the problem that Conceptual Engineering just changes the subject matter of debate.⁴⁴ Cappelen (2018) makes a strong case that Conceptual Engineers are not changing the subject-matter since subject-matters (or ‘topics’ as Cappelen calls them) are coarse-grained entities which can persist through long-standing debates (even when the words used may have undergone gradual change in meaning). In effect, Cappelen is proposing that there is a kind of common semantic ground which is revealed to us in the truth of the relevant same-saying reports. To make sense of this common-ground, we must acknowledge two kinds of semantic values: what is said (coarse-grained) vs what is meant (fine-grained).⁴⁵ So, two utterances of a single sentence may say the same thing, but may not mean the same thing, because the meaning of the sentence may be slightly different across the two times of utterance.

How does this help? It offers a compromise. The semantic commandment to not diminish the biblical word is arguably not violated by mere difference of meaning (and difference of use) – not when what is said can be preserved. In effect, this commandment is not, after all, (paradoxically) demanding that we do the impossible: it merely requires that our usage does enough to preserve what the Bible says. That’s not a complete resolution of all aspects of the puzzle of course, but it does go to show that insights drawn from the Conceptual Engineers can begin to provide the kind of solutions we might wish for with respect to paradoxes like the one considered here.

13. Interfaith dialogue

Those religious practitioners who promote inter-faith dialogue are keenly aware that the conditions for success are not easy – both from a practical perspective (of getting different faiths into the same room, speaking to each other) but also from a theoretical perspective: how can these two groups possibly communicate with, and understand, each other when they occupy such different perspectives?

The considerations of the previous section are highly germane to both issues because semantic common ground provides an important foundation for the possibility of inter-faith dialogue. There is a tendency in many contemporary social and philosophical debates to foreground issues of context and particularity. From a theoretical perspective it can even seem like a kind of radical contextualism is called for whereby there is no such thing as the meaning of a word, the truth-conditions of a sentence, the subject-matter of a debate, the issue at stake.⁴⁶ It can then seem that the participants to some important debate occupy hermetically sealed perspectives. No wonder the goal of genuine inter-faith dialogue can seem so quixotic. The idea of semantic common ground (via same-saying), however, provides an important corrective to these radical forms of contextualism; and, moreover, one which still does justice to context and particularity. That's because while the meaning of some word may be slightly different over time (and place, and speech community), what two speakers (who occupy different perspectives) say by using this word can nonetheless be the same. Such common ground thus provides a tool whereby we can recognise the semantic particularities of some religious community, but nonetheless also find ways to genuinely communicate and debate across the divide.

14. The name of god and coarse-grained aboutness

Perhaps the most important case study as regards the possibility of inter-faith dialogue (amongst the mono-theistic religions) concerns the name of God. With respect to the Abrahamic religions, the basic worry is that the meaning of ‘God’, the meaning of ‘Allah’, and the meaning of ‘Yahweh’ (or ‘Jehovah’) are all in fact different because each of the Abrahamic religions do not agree as to the essential properties of the divine creator.⁴⁷ Hence, they do not have the same concept of this creator; hence, the three names cannot refer to the same being.⁴⁸ Consequently, these religions are simply not worshipping the same god, despite the fact that they share a common ancestry through the figure of Abraham.

It might be thought that this issue can be (easily) resolved by attending to the various developments in the theory of meaning and reference that have occurred over the past 50 years. In particular, it might be thought that the so-called causal-historical theory of proper names (popularised by Kripke (1980)) provides a neat solution to the question as to whether the Abrahamic religions are all referring to the same being.⁴⁹ This theory says that the meaning of a proper name is not equivalent to (or even determined by) some associated set of descriptions of the form ‘The-so-and-so’. Rather, the meaning of a proper name is just its reference – where this reference is fixed by some initial baptism (or some functional equivalent of such a baptism). This baptism may take the form of an ostension: you point at something and say: ‘Let this be called Glasgow’. Or, the baptism may deploy a definite description to fix the reference: ‘Let this village be called: Glasgow’. If a subsequent use of the word ‘Glasgow’ bears the right causal-historical relation to such a baptism, then this ensures that this subsequent use refers to the same thing as the thing originally baptised. So, while a definite description may feature in the initial baptism by which a name gets its reference, such descriptions are not part of the meaning of the word. Indeed, someone may successfully refer to Glasgow using ‘Glasgow’ even though they associate with this name a cluster of definite descriptions – each of which is in fact false. (For example: the false claim ‘Glasgow is the main city on the East Coast of Scotland’.) This opens up the real possibility that the Abrahamic religions are, after all, referring to the same being, despite having non-equivalent concepts of this being.⁵⁰ That's because on this Kripkean kind of view, any individual concept associated with a proper name does not determine the reference of that name.

This kind of response, while highly promising, is not without its issues. For one thing, Kripke's famous modal argument against the descriptive view of names arguably does not get any proper purchase. That's because the canonical descriptions associated with God, by each of the Abrahamic religions, are necessarily satisfied by one and only one individual: God.⁵¹ So, for example, the following sentences have the same modal profile (if true then necessarily true; if false then necessarily false):

God is F. Allah is F. Yahweh is F.

The perfect being is F. The creator of the universe is F.⁵²

Indeed, once it is recognised that the Abrahamic religions agree on the canonical descriptions which are true of God then this opens up the way for a view of the reference of ‘God’/‘Allah’/‘Yahweh’ which is part descriptivist, part non-descriptivist.⁵³

Rather than try to settle the descriptivism/non-descriptivism debate over the name of God here, we want to float a different resolution of this issue which goes via the following promising hypothesis: if Cappelen (2018) is right that what is said is a coarse-grained semantic-value of sentences, then there ought to be sub-sentential semantic properties which are coarse-grained too. In effect, there must be a coarse-grained analogue of the semantic relation refers (for names) together with a coarse-grained version of the semantic relation is true of (for predicates). The evidence for such relations does not come about because of same-saying (indirect) speech reports as such but from same-aboutness reports. If we go back to the speech reports deployed in §10 as regards the two (temporally separated) utterances of the sentence ‘God is merciful’ then not only can we correctly report these utterances as saying the same thing, we can also correctly report them as being about the same thing. That is, ‘God’ (as used in the first utterance) is about the same thing as ‘God’ (as used in the second utterance). With suitable caveats (which ensures the sameness of the sentence uttered) the same goes for a range of relevant aboutness reports across place or perspective.

The upshot is that we can be neutral as to whether the reference of a proper name (as used in some sentence) has changed across place, time, perspective, while nonetheless holding that the proper name deployed is about the same thing. (Furthermore, this also allows us to be neutral as to whether some form of descriptivism about proper names is correct.) This then provides a further dimension to the idea of semantic common ground mooted in the previous two sections. Not only is this semantic common ground secured by the coarse-grained semantic values which Cappelen posits, namely, via what is said. It is also secured by the coarse-grained word-world relation: is about.

The practical benefit of the proposed resolution of this issue ought to be clear: when gathering together a group of theologians from the Abrahamic religions into a single conversation, we can justifiably announce: ‘we are all talking about the same god: Allah, Yahweh, God’.

15. Conceptual genealogy

Contemporary analytic theology explicitly includes an inherently genealogical dimension at its heart – via the systematic attention to the historical development of religious doctrine.⁵⁴ While analytic philosophy is not at all averse to the study of the history of philosophy, it is has typically been hostile to adopting a genealogical perspective whereby we (partly) analyse the nature of some thing or phenomenon by uncovering its historical development.⁵⁵ In this respect, analytic philosophy has, we contend, been overly narrow.⁵⁶

The Conceptual Engineering movement, meanwhile, turns out to be overtly genealogical in its methods. That ought to come as no surprise. When discovering that some design is faulty or need of improvement, the engineer is always keen to know why certain features of the design were originally introduced: what is their purpose or telos? Why did these features emerge in just this way? It is also common to find that Conceptual Engineers take concepts to have functions – which may well be evolutionary adaptations to their on-going usage across a variety of taxonomic and explanatory environments.⁵⁷ More generally, many Conceptual Engineers explicitly recommend re-engaging with the genealogical methods of Nietzsche and Foucault.⁵⁸ Take the concept of a person. Analytic philosophy, in its traditional guise, tells us that a person is a being with certain traits (sentience, or the potential for sentience; rationality or the potential for rationality) which has certain rights and values. This kind of analysis is considerably enriched when we consider the genealogy of this concept and the various complex roles it has played and was supposed to play.⁵⁹ Once this conceptual genealogy is in place, we find that the concept of a person has acquired multiple, overlapping, and often conflicting conceptual roles. It seems eminently ripe for some genealogically informed assessment to see if its meaning or function is still in good standing or is in need of improvement.

16. AI

A strong case can be made for applying the tools and methods of conceptual engineering to the new and challenging questions that AI poses for analytic theology.⁶⁰ With the exponential technological advancements in the field of AI, such systems are now exhibiting capacities and features that were once held to be exclusively human. Furthermore, such systems are even being described in terms which have largely been used with respect to humans, as a way of emphasising how human-like such AI systems are becoming.

This, therefore, poses a twofold problem for analytic theology. The first of these relates to the theoretical questions that must be answered as to whether or how AI is similar to humanity. But it is very evident that this can only be done when one examines what we mean by concepts such as intelligence, artificiality, thought, creativity, sentience and consciousness, amongst others. Such concepts might need to be replaced, changed or broken up in order to precisely capture that which is common to both humans and AI, and that which isn’t. Is the only difference between AI systems and ‘human systems’ the former’s artificiality, or is there something more? A promising way to provide a comprehensive answer to this question, however, is by revisiting how certain notions and concepts are understood and applying one or another of the various methodological treatments that have been explained above from the Conceptual Engineer’s repertoire.

But this is not all; the use of ‘human’ metaphors to describe AI systems also lead us to question how our humanity is conceptualised and understood (by, for example, thinking of the brain as our ‘software’ composed of a set of algorithms, and the body as ‘hardware’).⁶¹ The theologian is therefore compelled to re-examine certain theoretical core concepts that are fundamental to how we understand humanity, its place in the order of creation, and its relation to God; from what is special about our embodied nature (especially in light of the Christian notion of the Incarnation), to whether death now takes on a different definition, to how are we to articulate the concept of soul, and whether AI systems could ever be in possession of one!

However, analytic theology is not concerned only with that which is purely theoretical, but also with lived experience and the applied ramifications of the frameworks which are proposed. This, then, gives way to the second problem that analytic theology faces as one moves from the theoretical considerations to the practical applications of AI. These ‘applied’ considerations depend on however the theologian articulates the similarities and differences between humans and AI systems with respect to central core concepts such as the ones countenanced above. Which actions and activities can AI systems replicate in a meaningful and authentic way, as opposed to merely imitating without understanding? Would an AI system be able to assert its belief in an article of faith, or hope in the afterlife, or declare its undying love towards someone? How is sin and redemption to be understood for systems that carry out decisions based on pre-programmed algorithms? One might opine that this is unlike the free will that humans have, and yet others already question whether humans have free will in the first place, given the ‘pre-programmed’ nature of their biology and genes.

The above issues are only a small example of why the rapid progress in AI technology along with its ever-increasing pervasiveness in society should cause the theologian to re-examine the fortifications of its core concepts, and replace or rebuild where necessary.

17. Prospects

For reasons of space, we have left out three important themes which are also ripe for cooperation: context (which includes issues concerning particularism, relativism, uniqueness and intersectionality); ineffability (which includes the limits of thought, the limits of conceptual and semantic revision; the sublime; the divine; the nature of religious experience); and change (which includes progress, development, adaptation, but also subsumes issues as to whether spiritual progress requires conceptual progress; transformative experience; and the nature of (conceptual) revelation).

Furthermore, there are two kinds of engineering which are closely related to conceptual (or semantic) engineering which, we contend, will also prove to be of importance in strengthening the bridge being proposed. The first of these is: meta-semantic engineering. Meta-semantics, very roughly, is the study of the rules, conventions, and mechanisms via which a word acquires, retains and loses its meaning. (We touched on the meta-semantics for the name ‘God’ in §12 above.) While semantic engineering is all about designing (and potentially improving) the semantic properties of the primary representational devices, meta-semantic engineering is all about designing and improving the mechanisms and rules via which a representational device acquires, retains or loses, its semantic properties. On this score, it is an entirely open question whether the meta-semantics for the names of necessary existents (such as God) differs (or should differ) from the meta-semantics for the names of contingent existents (such as you). Getting clearer on such issues is an important dimension to enriching the working hypothesis of this article.

The second form of engineering of further interest is what Cappelen calls worldly engineering.⁶² On Cappelen's view, in engineering the meaning of the word ‘disability’ we also end up revising what it is to be disabled. The former is semantic revision; the latter is a kind of worldly revision – in particular, it involves changing the social fabric of our world. This insight provides an important antidote to the worry that Conceptual Engineering is just fiddling with the meanings of words – just an idle activity without any practical application. For this reason, this under-explored dimension to Conceptual Engineering provides a further area of potential cooperation – by way of applied philosophy and applied theology. These speculations, however, are topics to be explored on another day. For now, we hope to have made a good initial case that our working hypothesis is not only immune from worries as regards being too radical or too trivial, but also that it merits further exploration.

Disclosure statement

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

Notes

1 See §4 for Cappelen’s broader conception. Some philosophers prefer ‘Conceptual Ethics’ to ‘Conceptual Engineering’ e.g. Burgess and Plunkett (2013). We take these terms to be interchangeable.

2 As cited in Cappelen and Plunkett (2020, 1).

3 It may well be that traditional theology is more conservative with respect to core theological concepts but more open to conceptual revision/replacement for non-core concepts.

4 Another challenge is: Conceptual Engineering isn’t the right method for philosophy; so, it’s not a good method for analytic theology; so, our working hypothesis, while true, is useless. On this score, Conceptual Engineering faces three problems: (1) It is too utopian (see §4.5). (2) Conceptual Engineers are changing the subject (§12). (3) It is not needed (see Greenough n.d.).

5 See Cappelen (2018, 2020a, 2020b), Cappelen and Plunkett (2020).

6 Representational devices could be mental, physical or abstract; they could be fine-grained or coarse-grained; they could be immutable or plastic; they could be extensional, intensional or hyper-intensional; they could be internal states or partly external.

7 For example: ‘What is X?’, ‘How is X possible?’, ‘How should we solve this puzzle involving X?’. Conceptual Engineering could be a first-resort tool, a last-resort tool, an occasional tool or the only tool.

8 Reasons for improvement include: empty reference, lack of meaning, underspecificity, unclarity, ambiguity, incoherence, falsity, vagueness, lack of simplicity, lack of explanatory value, unknowability.

9 Cappelen (2018) doesn’t admit concepts. A concept-unfriendly version is: introduce a new surrogate word which has a new, better meaning.

10 Or, the primary representational devices are concepts, conceived of as the constituents of thought. (See Isaac 2020, 2021a, 2021b, 2021c, 2021d.)

11 Again, you could drop concepts and treat meanings as the primary representational devices.

12 Cf. Tarski (1944, 355).

13 Or they could be meanings and words.

14 See Burgess and Plunkett (2013, 1096). See also Cappelen (2018), Jorem (2021), Koch (2020), Queloz and Bieber (2021).

15 See Greenough ‘Utopian Philosophy’ (forthcoming) for this response to the Implementation Challenge.

16 There is some discussion as to whether analytic theology’s roots stretch much farther back to Scholasticism. Analytical Thomism is one such example (see Haldane 1997), though to delve further into this debate on the true roots of analytic theology would take us too far afield given our current aim.

17 See for example: Baker-Hytch (2016), Rutledge (2021), Torrance (2019).

18 For example Lebens and Segal (2022).

19 Some examples of this are found in Abdelnour (2023), Saemi and Davison (2020), Turner (2022).

20 This can be seen in various fields of study, from logic (Cotnoir 2019), to metaphysics (Rea 2020), to epistemology (see Macdonald (2014) for how this was crucial to analytic theology’s beginning).

21 From moral issues such as how accountability is a virtue (Torrance 2023), to social issues such as racism, oppression, and homophobia. (See Panchuk and Rea (2020).)

22 See Rodkey and Miller (2018) for a comprehensive survey of radical theology.

23 Schiffer (1996, 2003, 2004), Scharp (2013, 2021).

24 Cappelen (2018) defends this revivalist view. cf. Strawson (1959, 9).

25 Replacement forms of Conceptual Engineering tend to be futuristic; revisionary forms tend to be revivalist.

26 The requisite revisions or replacements may only be needed in certain contexts. Scharp (2013) allows that the concept of truth only needs replacing when doing truth-conditional semantics.

27 See (Schiffer 2003, 2004).

28 For example, mystery may be a positive feature. Cf. Anderson (2007, 2018).

29 Less misleading terminology is: ‘The ascending surrogate for truth’, ‘The descending surrogate for truth’.

30 Hence, the putative Orwellian dimension to this form of Conceptual Engineering.

31 Greenough (2019) argues that replacing one central concept entails replacing all the concepts connected to it – a kind of conceptual genocide.

32 So, on this view, Schiffer’s concept of justification* just is the concept of justification – only in modified form. As such, it doesn’t need a new name.

33 See e.g. Beall and Cotnoir (2017), Benton, John, and Hawthorne (2018), Cotnoir (2018, 2019); Fales (2009), Macdonald (2014), Visala (2016), Wood (2021a).

34 See Greenough (n.d.), for a discussion of the extensional version.

35 Exodus 4, 11: The Lord said to him, ‘Who gave human beings their mouths? Who makes them deaf or mute? Who gives them sight or makes them blind? Is it not I, the Lord?’

36 See Barnes (2016).

37 Because they have made assertions which are not known to be true.

38 Barnes (2016) does not explicitly suggest this definition but it accords with her general view. Whether she counts as a Conceptual Engineer is an open question for our purposes.

39 For a comprehensive historical development of this practice in different religious see: Olson (2007)

40 There is little philosophical discussion of the incel movement. See Melo Lopes (2023) for a feminist perspective that also surveys some of the present literature.

41 ‘Ye shall not add unto the word which I command you, neither shall ye diminish ought from it, that ye may keep the commandments of the Lord your God.’ Deut. 4:2. (Cf. Revelation 22:18–19; Proverbs 30:6.)

42 We are screening off the issue of translation here.

43 See Cappelen (2018, 31 ff.) drawing on Dorr and Hawthorne (2014) on meaning change paradoxes.

44 See Strawson (1963).

45 Cappelen endorses: the subject matter of the sentence S (used at time t₁) is the same subject matter of the sentence S (used at a distinct time t₂) if and only if these sentences (as used) say the same thing.

46 See Recanati (2002), Searle (1978), Travis (1985, 2008).

47 Ditto for the other Abrahamic religions (Druze, Yezidi and more).

48 Furthermore, since there is only one God, then the worry is that only one of these names refers to anything.

49 See e.g. Miller (1986), Alston (1988).

50 Cf. Putnam (1975).

51 Cf. Kaplan (1978)

52 Here the associated definite descriptions denote the same individual in all possible worlds.

53 Harris (1991) proposes that we return to cluster version of descriptivism for ‘God’.

54 See Hudson (2021), McCall (2021), Stump (2018).

55 See e.g. Popper (1957).

56 See Dutilh Novaes (2015), Ballarin (2022).

57 Carnap (1963), Thomasson (n.d.).

58 Plunkett (2016).

59 See LoLordo (2019), Taylor (1975, chap., 4).

60 For a broader overview of intersections between AI and theology see Oviedo (2022).

61 For more on the ’dual direction’ of these types of metaphors, see: O’Gieblyn (2021).

62 Cappelen 2018 (chap. 12). See also Greenough (n.d.) for an account of Reality Engineering.