Linguistic innovation for gender terms

ABSTRACT

This paper explores the political prospects of linguistic innovation for gender terms. It develops and defends a model that leaves the concept of woman in place but replaces the term with a novel one that lacks the problematic social and political connotations. It is shown that linguistic innovation reaps all the political benefits of competing proposals in conceptual ethics, without any of the downsides.

KEYWORDS:

• Conceptual engineering
• conceptual ethics
• feminist philosophy
• gender
• race

1. Introduction

Can we improve the world by engineering better representational devices? This is perhaps the key question in the theory of conceptual engineering. While it continues to be hotly debated (e.g. Carnap 1950; Strawson 1963; Capellen 2018; Simion and Kelp 2020), we will set it aside for the purposes of this paper and simply assume that the answer is yes. What we will focus on instead is the practice of conceptual engineering, which has come to enjoy an increasing degree of popularity in a wide range of philosophical research. There are a number of live proposals for engineering representational devices, including the concepts of truth (e.g. Scharp 2013), freedom (e.g. van Inwagen 2008) and belief (e.g. Clark and Chalmers 1998). All of them are aimed at making the world a better place by improving our representational devices.

In this paper, we will explore the prospects of engineering a key concept in feminist philosophy, to wit, the concept of woman. Many engineering projects, including the examples just mentioned, have been motivated by semantic defects of the target concepts such as inconsistency. In contrast, the main motivation for engineering the concept of woman is political. Women have historically been subject to various forms of subordination. Since one is ‘marked’ for application of oppressive pressures by one’s membership in some group or category (Frye 1983), and since, in turn, one’s membership in some group or category is marked by the application of the corresponding concept, a key tool in the hand of the oppressors has been the concept of woman itself. A key aim, then, of engineering the concept of woman is to do away with this injustice and to alleviate harm brought upon women as a result.

While there may well be a variety of dimensions along which we might consider engineering the concept of woman in order to achieve political justice, the perhaps most influential strategy originated with Sally Haslanger’s influential (2000) paper. Our central aim is to develop an alternative to this proposal that compares favourably with it.¹ In contrast with Haslanger, we will achieve the aims of her engineering project through linguistic innovation rather than conceptual engineering.

Here is a game plan for the paper. Section 2 outlines Haslanger’s engineering project and its political ambitions. In Section 3, we will raise a number of worries for Haslanger, some old and some new. Finally, in Section 4, we will develop an alternative to Haslanger and argue that it can secure a number of benefits of Haslanger’s view whilst avoiding the problems she encounters.

2. Haslanger’s engineering project

Haslanger’s engineering project is for gender and race concepts. Since our interest is primarily with her proposal for engineering the concept of woman, that’s what we will be focusing on here. That said, there is reason to believe that a number of the points that we will be making carry over to Haslanger’s proposal for race concepts as well.

To begin with, Haslanger is pretty clear about the political ambitions of her project when she writes: ‘the primary goal is an analysis of gender that will serve as a tool in the quest for sexual justice’ (2000, 37). In other words, she wants an analysis of gender that is conducive to bringing about justice. According to Haslanger, our concept of ‘woman’ is in need of work; the concept in use carries politically problematic connotations: historically, it came to be associated with social and political subordination. Haslanger proposes to engineer ‘woman’ such as to bring these connotations into clear view; the key move is to engineer the concept of woman so as to feature an explicit subordination condition. Here is her specific proposal:

S is a woman iff_df S is systematically subordinated along some dimension (economic, political, legal, social, etc.), and S is ‘marked’ as a target for this treatment by observed or imagined bodily features presumed to be evidence of a female’s biological role in reproduction. (2000, 39)

According to Haslanger, her redefinition of the concept of woman can be an effective tool towards bringing about gender justice. The key thought is to work in the direction of justice by engineering subordination in virtue of certain bodily features into the definition of the concept of woman. In this way, the target concept is meant to become a ‘negative ideal’ (2000, 46), which is conducive to bringing about the much-needed justice. As she sees it, the ultimate goal of her project is to get to a point at which there aren’t any women anymore because the extension of ‘woman’ is empty (2000, 46). In this way, Haslanger’s project is abolitionist.

3. Worries for Haslanger

3.1. A source of positive value

One worry one might have about Haslanger’s project is that the success of the abolitionist strategy might encounter resistance form a non-negligible number of individuals in virtue of their desire to remain women. Mari Mikkola (2011) was, to our knowledge, the first to express scepticism towards abolitionism for this reason: according to Mikkola, the feminist concept and the everyday conception of gender categories come apart: according to a very popular view in feminist philosophy, gender and sex are ontologically different categories, in that the former is socially constructed, while the latter is biological. In contrast, the folk conception fails to distinguish the two. In other words, both feminists and laypeople may well agree that our social norms are systematically disadvantaging women. However, while feminists think that the problematic norms are constitutive of the category, laypeople typically take them to be merely regulative. As a result, Mikkola argues, it would be unsurprising if the abolitionist feminist project were to encounter resistance from laypeople. After all, one may wonder, why pursue abolitionist aims rather than target the bad regulative social norms that harm women? In this way, there is reason to be pessimistic that the abolitionist project will have traction with laypeople (2011, 75).

Furthermore, Mikkola argues, since the vast majority of the general population takes gender categories to not be constitutively problematic, but rather merely subject to harmful social regulation, it is unsurprising that we often find people who take themselves to enjoy and even take pride in being a member of the gender category in question. Of course, if this is so, we should expect even further resistance to the abolitionist project (2011, 75).

If Mikkola is right, we should expect, on empirical grounds, that implementing the negative-ideal based abolitionist ambitions will, indeed, be harder than we initially thought.

3.2. Trans inclusivity

Another important objection to Haslanger’s account is that it is insufficiently trans-inclusive. Here is one compelling argument for this claim, due to Katharine Jenkins. Not all trans women have or are imagined to have bodily features presumed to be evidence of a female’s biological role in reproduction. A fortiori they cannot be marked for subordination along some dimension in virtue of this. Haslanger’s account thus misclassifies at least some trans women as non-women.

Moreover, Jenkins is clear that Haslanger’s abolitionism is at loggerheads with trans inclusivity. Here is Jenkins:

[…] the invitation to give up identifying with the term ‘woman’ will tend to have a different meaning for cis women, who are pushed toward this label, than for trans women, who typically face a struggle to claim it. Giving up one’s attachment to thinking of oneself as a woman may feel very different if one’s right to identify in that way has received at best a partial or patchy recognition. Thus, even if ceasing to identify with the label ‘woman’ may be a constructive option for some women, it will very probably be undesirable or unworkable for many others, including at least some trans women (417–418).

Jenkins’s key move is to engineer two concepts of woman. The first, which she labels ‘woman as class’, corresponds to Haslanger’s definition of the concept of woman. The second, ‘woman as identity’, is defined as follows:

S has a female gender identity iff S’s internal ‘map’ is formed to guide someone classed as a woman through the social or material realities that are, in that context, characteristic of women as a class. (Jenkins 2016, 410)

The central idea is that the concept of woman-as-identity is a concept of gender to be countenanced alongside the concept of woman-as-class such that both have equal theoretical status. It is easy to see that this promises to avoid the charge of insufficient trans-inclusiveness. While Jenkins acknowledges that some trans women will not be classified as women by the concept of woman-as-class, all trans women will be classified as women by the concept of woman-as-identity. Moreover, Jenkins argues that every successful engineering project for the concept of woman must countenance both the concept of woman-as-class and of woman-as-identity, and place them on equal footing. A failure to do so will be tantamount to a failure to properly understand some forms of gender-based injustice.

We do not mean to deny that Jenkins’s proposal has the resources to solve problem of trans inclusivity that Haslanger encountered. That said, we worry that the prospects of addressing both forms of gender-based injustices by countenancing two concepts and placing them on equal footing faces serious obstacles. In particular, we think that once we take on board Jenkins’s proposal, the prospects of Haslanger’s abolitionist project for ‘woman-as-class’ are dim.²

To see this, note that one important decision that needs to be made on this picture is: which concept will the term ‘woman’ express: the concept of woman-as-class or the concept of woman-as-identity? As we are about to argue, to do the work against relevant form of gender injustice in the way envisaged, each concept needs the association with the term. That is to say, to achieve gender justice in the abolitionist way envisaged by Haslanger, the term ‘woman’ must express the concept of woman-as-class. At the same time, to achieve trans inclusivity, the term ‘woman’ must express the concept of woman-as-identity. We thus end up with a dilemma. Let’s look at the details.

To begin with, as Jenkins also realises, to solve the problem with trans-inclusivity that Haslanger encounters, it is imperative that the term ‘woman’ goes with the concept of woman-as-identity (417–418). The trouble is that it is key to the very motivation for engineering the concept of woman-as-class that the term ‘woman’ expresses the concept of woman-as-class. To see this, note that it is hard to see why we should engineer the concept of woman-as-class unless we have abolitionist ambitions. What would be the point of enshrining subordination of women-as-class into the very definition of the concept of woman-as-class unless we had the ambition to abolish women-as-class? As we have seen in Section 2.4 above, there is even reason to worry that engineering the concept of woman-as-class without abolitionist ambitions will not only not be of much use, but, to the contrary, is outright dangerous in that it might lead to further mistreatment of women.

Once we are clear that abolitionist ambitions are key to any project that ventures to engineer the concept of woman-as-class, we can see why it’s imperative that the term ‘woman’ expresses the concept of woman-as-class. Since the concept employed in the bulk of ordinary thought and talk about women is the concept expressed by ‘woman’, if ‘woman’ expresses the concept of woman-as-class, the negative ideal will be constantly on our minds.³ In this way, if ‘woman’ expresses the concept of woman-as-class, Haslanger’s engineering proposal offers an effective means towards achieving the abolitionist aims that are part and parcel of any project that ventures to engineer the concept of woman-as-class. In fact, relative to the Haslanger’s design plan for her engineering project, this means it is essential for achieving the abolitionist ambitions.

In contrast, if we tie the term ‘woman’ to the concept of woman-as-identity, what will be constantly on our minds in ordinary thought and talk about women is the concept of woman-as-identity, not the concept of woman-as-class. Crucially, however, the concept of woman-as-identity does not come with a negative ideal. This means that the negative ideal will not be constantly on our minds. As a result, we will inherit the abolitionist ambitions that come with engineering the concept of woman-as-class whilst, at the same time, jeopardising Haslanger’s effective means towards achieving them and abandoning what, relative to the design plan, is an essential such means. It is easy enough to see that this undermines the very motivations for engineering the concept of woman-as-class in the first place.

But couldn’t we be somewhat flexible about the terms? Couldn’t we allow that one of the concepts ends up going with a different term? In particular, couldn’t we allow that the concept of woman-as-class ends up going with a different term? Unfortunately, there is reason to think that there is less room for flexibility. This is because the term ‘woman’ is a key component in any project that aims engineer the concept of woman. After all, the term ‘woman’ is the label that helps us language users recognise that what the concept at issue is the concept of woman rather than a different concept. Most importantly for present purposes, the term ‘woman’ is what gives us epistemic access to the fact that the engineering proposal in question targets the concept of woman rather than a different concept. It’s not as if we had only rhetorical advantages of using the term ‘woman’. If we want to engineer the concept of woman, we cannot do so without it.⁴

One might wonder whether the best option isn’t simply to abandon the concept of woman-as-class and hold on to the concept of woman-as-identity. Perhaps. That said, at the very least, this move will mean a lost opportunity. One of the sources of injustices against women is the very use of the term ‘woman’ and the associated concept of woman. They already perpetuate injustices against women, for instance by perpetuating a world view of which these injustices are part and parcel. One attractive feature of Haslanger’s proposal is that it promises to work against this source of injustices by changing the concept of woman expressed by the term ‘woman’ such that it incorporates a negative ideal. In contrast, the concept of woman-as-identity doesn’t. What’s more, there is nothing in the concept of woman-as-identity that provides reason to believe that using this concept rather than the old one will allow us to alleviate the injustices perpetuated against women by the very use of the term ‘woman’ and the concept it expresses that Haslanger’s proposal promises to allow us to make. For instance, there is no reason to believe that using the concept of woman-as-identity instead of the old concept of woman will work against perpetuating a world view of which injustices against women are part and parcel. On the contrary, there is every reason to think that the use of the concept of woman-as-identity will continue to perpetuate the very same unjust world view but in such a way that everyone who falls in the extension of woman-as-identity is affected.

These two important worries aren’t the end of the problems that Haslanger’s proposal faces. In what follows, we will add two further difficulties to the list.

3.3. A game no one plays

One of the potential advantages of engineering the concept of woman is to counteract the perpetuation of the injustice by altering our way of thinking about the world. It might be thought that Haslanger’s proposal will do particularly well on this front. After all, once we have adopted the concept of woman-as-class, the negative ideal will be constantly on our minds, i.e. whenever we deploy the engineered concept. And isn’t it hard to think of a better way for the negative ideal to take effect than by being constantly on our minds?

Unfortunately, there is reason to think that the prospects of Haslanger’s proposal are much less bright than they may appear at first sight. In a nutshell, the reason for this is even if we agree on the proposed new definition, it is plausible that it is the old concept that continues to get used in ordinary life. If so, whatever harm is done by employing the old concept (perpetuating the unjust world view, etc.) continues to be done, even if we were able to agree on the new definition.

Let’s look at the details. Note first that even if a concept has a definition, one can be competent with that concept even if one does not know this definition. In fact, one can be competent with the relevant concept even though one has no idea that the relevant concept admits of definition in the first place. By way of example, think of the concept of drake. While this concept admits of a neat definition in terms of the concepts of male and duck, it is hard to deny that anyone who can reliably identify drakes by their distinctive look, who will reliably enough apply the term ‘drake’ to drakes and drakes only, etc. must be credited with a concept of drake. After all, any such person will be able to come to believe propositions about drakes such as that they are looking at a drake, which is something they can’t do that unless they have the concept of drake.⁵ But of course, one may well be able to do this without knowing the definition of the concept of drake. In fact, one may even be unaware of the fact that the concept of drake admits of definition in the first place.

Suppose, then, that the ordinary concept of woman has a definition. Even so, one may be competent with that concept even though one neither knows this definition nor even knows that the concept of woman admits of definition in the first place. After all, here too, it is hard to deny that anyone who can reliably identify women in various ways (by looks, names, etc.) and reliable enough applies the term ‘woman’ to women and women only must be credited with the concept of woman. Again, one may be able to do this without knowing the definition of the concept of woman or even knowing that the concept admits of definition.

Second, even if a concept has a definition, the definition need not always be deployed when we use the relevant concept, not even when it comes to classifying objects as falling under the concept. Consider the case of the concept of drake once more. Of course, we may classify certain objects as drakes by ascertaining whether they are both ducks and male. However, we can also classify them as drakes based on their looks. In fact, it may well be that the definition is rarely if ever used in ordinary thought and talk. Again, the concept of drake serves to illustrate this nicely. Most do not deploy the definition of the concept of drake in ordinary thought and talk, not even when it comes to classifications of things as drakes. On the contrary, they will typically classify things as drakes based on the look of drakes, rather than by ascertaining whether the object is both a duck and male. This is no surprise given how unwieldy it is to do so by ascertaining that the conditions for the concept of drake (especially being male) are satisfied.

Likewise, even if the ordinary concept of woman has a definition, this definition need not be employed in ordinary use, not even when it comes to classifying certain people as women. In fact, in the case of the concept of woman, it may well be that we do not deploy the definition in ordinary thought and talk, not even when it comes to classifying people as women. Rather, here too, we will typically classify people as women based on their looks, names, etc.⁶

Now, suppose we agreed to replace the ordinary definition of woman with Haslanger’s definition or, in case the concept of woman does not have a definition, to inaugurate Haslanger’s definition. What difference will this make to our ordinary thought and talk? The overwhelmingly plausible answer is: quite little. The old concept of woman will remain in use in much ordinary thought and talk. There are at least two reasons for this. First, Haslanger’s definition is too impractical to feature much in our ordinary classifications of people as women. After all, just as in the case of the concept of drake, it is so unwieldy to first ascertain that the conditions for Haslanger’s definition of the concept of woman (especially the subordination condition) are satisfied.

Second, the bulk of our classifications of people as women proceeds automatically and in a way that is, paradigmatically, not under our voluntary control. Walking down a busy street, you will classify lots of people as women without even giving it conscious thought. Seeing a certain name written on the door of an office, you will classify its owner as a woman, etc. What’s more, this, paradigmatically, happens automatically, and in the vast majority of everyday cases there is little one can do to prevent it from happening even if one wanted to. And when you do so, you use the old concept of woman.

That is not to say, of course, that one has no control whatsoever over concept application: indeed, it is plausible to think that, insofar as one can successfully overcome all kinds of well-entrenched cognitive biases, one can also work on concept replacement. Rather, the problem is that (1) de-biasing work is notably hard, and comes with mixed success, and (2) going back to the point before, insofar as one can be conceptually competent with a concept C without even knowing the definition of C – indeed, without even knowing that it admits of a definition – it seems plausible that a significant number of competent employers of ‘woman’ will remian unaffected in their conceptual practices by Haslanger’s project.

Furthermore, even if it turns out that we can educate our conceptual employment practices (in the way in which bias training (tries to) correct biases), Haslanger’s definition will still not do the trick: after all, the extension is very minimally changed by her engineering project. What is supposed to do the work is the explicit employment of the problematic definition: via explicitly employing it, the thought goes, we will become more aware of how problematic it is, and the concept of woman disappears altogether. Non-explicit employment will not do the trick, because the problematic feature will not be in front of our minds at the employment moment, and thus it is unlikely to result in concept abandonment.⁷

If this isn’t immediately obvious, consider once more the concept of drake. Suppose you have just deployed the concept of drake, say because you have classified a certain animal as a drake by perceptual means. Suppose you are also competent with the definition of the concept of drake. If so, we may expect you to accept the inference from the proposition that the animal is a drake to the proposition that it is male as entirely obvious. In particular, we may expect that the question as to whether the animal is male is no longer open for you, even if you have classified the animal as a drake by perceptual means. Suppose next, and analogously, that you have just deployed the concept of woman, say because you have classified a certain person as a woman by perceptual means. Suppose you are also competent with Haslanger’s definition. If it is really the concept that is defined in the way Haslanger suggests that you deployed, then we may expect you to accept the inference from the proposition that the person is woman to the proposition that she is subordinated as entirely obvious as well. And, again, in particular we may expect that the question as to whether this person is subordinated is no longer open for you, even if you have classified the person as a woman by ordinary (e.g. perceptual) means. But this, we submit, is exactly not what we’d expect to find. On the contrary, when you have classified a person as a woman by ordinary (e.g. perceptual) means, we’d expect the question as to whether she is subordinated to remain open for you. Since had you deployed Haslanger’s engineered concept, we’d expect the question not to have remained open for you, there is reason to believe that you didn’t deploy Haslanger’s engineered concept. By the same token, there is reason to believe that you deployed the old concept instead.

To be clear, we do not mean to say that, in the kind of situation we are imagining (in which the Haslanger’s definition of the concept of woman has gained widespread acceptance), you never use Haslanger’s engineered concept. On the contrary, in borderline cases, for instance, you may very well use the definition to classify a certain person as a woman. Rather, our point is that there is a wide range of cases in which you will simply not use Haslanger’s engineered concept but rather the old concept – whatever the latter may involve. And this is where the problem for Haslanger rears its head. First, since the old concept continues to be in circulation, the project of transforming the old concept into Haslanger’s engineered concept has arguably failed. Second, since the very use of the old concept of woman plausibly contributes to perpetuating injustices against women (as per our initial assumption) – for instance, by perpetuating a world view of which these injustices are part and parcel – to the extent that we continue to use the old concept of woman, we continue to perpetuate these injustices against women. Since there is reason to believe that even if Haslanger’s definition of the concept of woman were to be widely accepted, the old concept would remain in use in much ordinary thought and talk, the problems that use of the old concept of woman brings in its train continue to remain in place also.

3.4. A dangerous game

Our previous worry granted Haslanger that it is hard to think of a better way for the negative ideal to have positive effect than by being constantly on our minds. However, it challenged the thought that, as a result of her engineering project, it would indeed be constantly on our minds. The second worry is even more pessimistic. It challenges Haslanger’s project on normative grounds. In a nutshell, the worry is that the negative ideal might have a negative rather than a positive effect. Let us explain.

Even if we grant that engineering the concept of woman to become a negative ideal will work in societies in which it is sufficiently well recognised that no one ought to be subordinated in virtue of certain bodily features, it is not clear that it will work in societies in which recognition of this fact is not widespread enough. And, of course, we are not even mentioning societies in which women are widely enough believed to be inferior. The worry is that, rather than challenging male dominance, the concept of woman-as-class will play into the hands of male dominance in those societies, for instance, by opening up new avenues of ‘justifying’ various forms of mistreatment of women.

One might worry that this is unfair to Haslanger. After all, isn’t it the case that pretty much any definition can be put to nefarious use in the wrong hands? If so, pointing out that the concept of woman-as-class can be put to nefarious use carries little traction against Haslanger.

We don’t mean to deny that pretty much any definition can be put to nefarious use. That said, there are differences in the negative potentials between definitions. And we worry that definitions that enshrine negative values (such as subordination) into the very fabric of a certain concept have a particularly high potential for negative impact. To see this, consider the perhaps most horrific attempt at conceptual engineering in human history. It centrally featured the concept of subhuman which gained currency in Nazi Germany. This concept had negative values, such as, among others, the subordination of ‘non-Aryan’ people enshrined into its very meaning. It was a particularly powerful tool for the Nazi regime because it helped the regime ‘justify’ its atrocities. And it also helped those contributing to their perpetration overcome natural psychological barriers to inflicting this kind of treatment to fellow humans.

The case of the concept of subhuman is a sobering reminder of how enshrining the subordination of certain minorities into the very meaning of a concept can make a substantive contribution to furthering the oppression of these minorities. Even if pretty much any definition can be put to nefarious use in the wrong hands, the historical precedent provides reason for thinking that the concept of gender as class is has a particularly high potential for damage.

4. Linguistic innovation

Haslanger’s engineering proposal faces a number of worries. Can we do better? We believe that the answer to this question is yes. In what follows, we will briefly sketch an alternative to Haslanger’s proposal and will show that it holds out the hope of avoiding the difficulties it encounters.

4.1. She’s a ‘She’

Rather than aiming to engineer the concept of woman, our proposal begins with linguistic innovation. It brings to life a new linguistic practice. In particular, we want to suggest targeted use of an existing variation in terminology: ‘she’ used as a noun, such that ‘she’ applies to x if and only if ‘woman’ applies to x: x is a she just in case x is a woman. While by itself this may do little to make the world a better place for women, there is a second component of the view. This is to adopt a social norm that prohibits using the word ‘woman’ in recognition of the injustices that have been inflicted upon women throughout history and encourages using ‘she’ instead in recognition of the justice they deserve.⁸

What are the benefits of this proposal? Recall that Haslanger’s main motivation for her engineering project is political. Women have historically been subject to various forms of subordination. A key aim of engineering the concept of woman is to do away with this injustice and to alleviate harm brought upon women as a result. Can the present proposal achieve this?

First, it is important to note that any (conceptual or linguistic) engineering proposal that is aimed at making progress towards removing the problematic connotations associated with a particular representational device will face a normative conflict,⁹ between political reward and cost of implementation: on one hand, for ease of implementation, the engineer needs a term that everyone will recognise as co-extensive with the term at issue – in our case, ‘woman’. On the other hand, in any term that has long been used and is co-extensive with ‘woman’ risks having a similar history of subordination, and thus be less efficient than a completely new term in getting rid of the problematic connotations.

While we think that this normative conflict is not easy to overcome, we believe that it is up to the engineer and their community (see fn. 7) to identify the proposal that achieves the best balance of implementation cost and political benefit. These considerations have guided our particular proposal in this case, but we are very open to better proposals, that achieve a better balance in this regard.

There is further reason to think that our proposal can help us move in the right direction. To see how, recall first that the strategy of enshrining subordination into the very meaning of a concept has historical precedent that provides cause for concern. The present strategy also has historical precedent. However, it is one that, if anything, provides some reason for optimism. Since the historical precedent allows us to see the potential benefits of the present strategy more clearly, we will start with it.

The historical precedent we have in mind is case of ‘negro’.¹⁰ Linguistic innovation including the targeted use of an existing variation in terminology in conjunction with a social norm prohibiting the use of ‘negro’ in recognition of its association with a history of subordination and degradation of black people have decreased the use of ‘negro’ across wide segments of the population.

Now, we don’t mean to say that this has solved the problem of racism in our society. However, it does constitute an improvement, in a number of ways.

First, if you want to rectify an injustice that has been done, recognition the fact that an in injustice has been perpetuated is an important step. The linguistic innovation in question offers a way of taking it.

Second, it brings a variety of prospective social penalties in its train. For instance, in the case of ‘negro’, thanks to the social norm, politicians, newspapers, etc. systematically using this term to refer to black people will not have a bright future, at least in mainstream politics, news reporting, and so on. Similarly, the social norm may bring about ostracism of users of ‘negro’. If you persist in using ‘negro’ to talk about black people, we certainly won’t associate with you for much longer.

Even so, one might wonder whether this really helps counteracting the perpetuation of an unjust worldview which takes black people to be inferior. After all, even if we change our way of talking about black people, our attitudes towards them may remain unchanged. But clearly that’s enough for the unjust worldview to live on. Even worse, this world view will continue to be perpetuated in much the same way as before, with ‘black’ playing the role of ‘negro’. In this way, there is reason to think that linguistic innovation doesn’t help much on this front. What is needed is a change in people’s attitudes. But that’s not something that linguistic innovation can deliver.

Again, we don’t mean to suggest that the kind of linguistic innovation that we are considering will bring about a speedy and large-scale change in people’s attitudes. At the same time, we think that there is some reason for optimism that it can contribute even towards a change in people’s attitudes, if only in the long run.

To see this, note that the proposed linguistic innovation arguably makes an important contribution towards generating conditions that are conducive to a change in attitudes. For instance, once we have adopted a social norm that prohibits use of the term ‘negro’, hearers have reason not to respond to assertions featuring the term ‘negro’ with belief. After all, once the social norm is in place, the chance that assertions featuring this term have contents that are either false or unjustified goes up dramatically. But, of course, if hearers do not respond to assertions featuring the term ‘negro’ with belief, one key mechanism for perpetuating the unjust worldview will be blunted considerably.

Moreover, speaking in ways that recognise those who have suffered injustices as having suffered them and as deserving justice gives us additional opportunities to identify and challenge manifestations of racist attitudes. After all, when we encounter these manifestations, say in an assertion of a racist generic about black people, we just may come to wonder whether what’s being said here isn’t one of the injustices that black people have suffered.¹¹ And we may challenge the assertion accordingly, for instance, by saying: ‘What’s your reason for thinking this?’ or ‘That really just sounds like a stereotype about black people.’ These challenges will prevent the assertion from entering the common ground of the conversation. And even if they don’t change the speaker’s mind immediately, continued challenges and criticisms of this sort may have a better an effect in the long run. More importantly yet, these challenges and criticisms will prevent the perpetuation of this attitude among hearers. For instance, a third party who would have responded with belief to the assertion had the assertion remained unchallenged, will be less likely to do so, now that a challenge has been presented. In this way, even if people’s attitudes won’t change with the inauguration of a new linguistic practice immediately, the new linguistic practice may contribute towards a gradual change of attitudes in the long run.

What comes to light, then, is that, while linguistic innovation is surefire cure for racism, it does allow us to achieve moral progress towards racial justice, for instance by providing important resources to counteract the perpetuation of a worldview in which black people are taken to be inferior. And, of course, of central importance for present purposes is that the parallel treatment we have proposed for ‘woman’ promises to allow us to make moral progress towards sexual justice as well, by providing important resources to counteract the perpetuation of a worldview in which women people are taken to be inferior, along similar dimensions.¹²

4.2. Back to the problems

With the benefits in play, let’s return to the worries that Haslanger encountered. First, it is easy enough to see that Mikkola’s worry about viewing gender as a source of positive value and the resistance to Haslanger’s abolitionist ambitions that result from this can be circumvented. After all, on the present proposal, the concept of woman and the social category that it constitutes remain unchanged. They can thus continue to be a source of positive value.

Second, the present proposal can also avoid the charge of being insufficiently trans inclusive.¹³ Note that ‘she’ was introduced not by definition but as a novel way of talking about women. As such, the extension remains unchanged; there is no need to ensure trans inclusivity through additional theoretical resources.¹⁴

Third, the present proposal allows us to make progress on the worry that, on Haslanger proposal, the old concept of woman will still be in use in much ordinary thought and talk, thereby perpetuating an unjust worldview. Of course, on the present proposal, the old concept of woman will continue to be in use. After all, the view has it that the new term, ‘she,’ still expresses the old concept of woman. At the same time, there are important differences. We provided reason to think that Haslanger’s engineering project will fail. In a nutshell, the reason for this is that it is not up to us what our concepts mean. In the case of the concept of woman, we have seen that there is reason to think that even if we all agree on Haslanger’s definition of the concept of woman, the old concept will remain in use unchanged. But, of course, if the engineering project fails in this way, the key engine of progress, to wit, the negative ideal, won’t get to work in the way envisaged either.

One important difference with our proposal is that, although what our concepts mean is not up to us, what words we use is up to us. In particular, nothing stands in the way of adopting the linguistic practice we proposed and to start using ‘she’ when talking about women. We have provided reason to think that doing so will allow us to make moral progress, to wit, by providing resources to counteract the perpetuation of an unjust worldview. While there are structural reasons for thinking that the engine of progress won’t work in the way envisaged for Haslanger, there aren’t parallel structural reasons that hampering progress for our proposal. In this way, our proposal promises to deliver a key improvement on this worry.¹⁵

Fourth, what about the worry that Haslanger’s concept of woman-as-class might play into the hands of male dominance, for instance, by ‘justifying’ new avenues of mistreatment of women in sexist societies? It is hard to see how the present proposal might do so. After all, the new linguistic practice we propose essentially features a social norm that recognises not only the injustices that that have been inflicted upon women throughout history and but also the fact that they are deserving of justice. And while, as we already mentioned, many good things can be put to nefarious use in the wrong hands, it’s not easy to see how this might be done in the present case.

In fact, there is reason to believe that the present proposal can be and often is a good thing even in oppressive societies. Consider, for instance, a highly oppressive society in which it is part and parcel of the accepted world view that women are inferior. The linguistic innovation we suggest provides members of this society with an avenue towards resistance against the oppressing regime and the world view it promulgates. If this isn’t immediately obvious, suppose for a moment, that you are living in Germany under the Nazi-regime, which, suppose, just recently adopted a linguistic practice of classifying Jews as ‘subhuman’. One way for you to resist the regime and the world view it promulgates is by refusing to talk and think of Jews as subhuman. It does without saying that resisting the Nazi-regime in this way is a good thing. Now, resistance action comes in two flavours: negative and positive resistance action.¹⁶ We can intentionally omit to act in a way that favours the regime, but we can also intentionally act in ways that blocks the regime and its tools of oppression. We find it plausible that the same goes for our conceptual engagements: what you are achieving here by refusing to take up a conceptual and linguistic innovation that feeds into a problematic regime can also be achieved by taking up a linguistic innovation that resists oppressive language – at least when the existing linguistic practice is problematic. After all, both linguistic moves are resistance moves: in virtue of refusing a problematic practice/adopting a reactionary practice, one is, de facto, involved in political resistance via linguistic means. One can resist a problematic system by refusing to engage in its practices, or, alternatively, by engaging in practices that speak against the system’s values. If so, then adopting the linguistic innovation we propose in the kind of sexist society under consideration is a good thing also.

4.3. Conceptual change?

Thus far we have highlighted the benefits of a certain kind of linguistic innovation, which leaves the concept of woman in place. Now we want to return to the concept once more. We’d like to emphasise that we do not aim to pursue a conceptual engineering project here, for reasons that will become clear momentarily. Rather, what we would like to explore is the possibility of a certain kind of conceptual change that results from the linguistic innovation. If the kind of linguistic innovation project we have proposed is successful, ‘woman’ may become disused. In fact, disuse is a good outcome. In recognition of the injustices that have been done to women in the past, we no longer use ‘woman’ (at least unless and until it is reclaimed).

That said, it is by no means a foregone conclusion that this is how things are going to play out. To see this, note that, on the present proposal, the use of ‘woman’ comes under selective pressure. However, rather than ending up disused, the use of ‘woman’ might instead adapt. And while there are many ways in which ‘woman’ might adapt, here we would like to focus on one, which takes its cue once more from the historical precedent we have been looking at, i.e. ‘negro’.

‘Negro’ was once a neutral classifying term (Hom 2010). Once it lost this status, it did not go out of use. Rather, it adapted in response to selective pressure and became a slur term. Now, what exactly this means for the concept expressed by ‘negro’ remains an open question, to be settled in the theory of slurs. However, there is a popular kind of view according to which the difference between a slur term and a neutral classifying term is a difference in meaning (e.g. Hom 2008). If this kind of view is correct, what happened to the term ‘negro’ is a shift in meaning and, by the same token, a conceptual shift; the latter, of course, need not amount to full concept replacement; the shift in the concept ‘negro’ expresses can merely undergo conceptual change. If ‘woman’ turns out to behave sufficiently like ‘negro’ by becoming a slur term, we will have a change in the concept expressed by ‘woman’ as well. In this way, the linguistic innovation strategy might still lead to conceptual change in a more indirect way.¹⁷

While this result would be in many ways worse than that the alternative in which ‘woman’ is simply disused, it is worth noting that it achieves at least one of the central aims of Haslanger’s project: if this is how things play out, there will no longer be any women.

Disclosure statement

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

Additional information

Funding

This work was supported by Horizon 2020 Framework Programme: [Grant Number 948356].

Notes

1 To the best of our knowledge, what we try to achieve here is either compatible with the bulk of the engineering projects that have been proposed in recent feminist literature or, insofar as they adopt Haslanger’s proposal, applies equally to them. That’s why we will not discuss them in any detail here. Rather, we will rest content with acknowledging that our proposal is best viewed as part of a wider project and by noting that there is a lot of exciting work that is being done on the issue as we speak (e.g. Bettcher 2013; Jenkins 2016; Mikkola 2011; Simion 2017, 2018).

2 Compatibly, Jenkins’s proposal may succeed in bringing about sexual justice on both fronts via a different, non-abolitionist route (thanks to XXX for pointing this out). We don’t mean to deny this, and we think it is an option worth exploring

3 At least provided that we bracket the worry from Section 2.3.

4 Couldn’t we maintain that the term ‘woman’ is polysemous and, as a result, expresses the concept of woman-as-class in some context and the concept of woman-as-identity in others?

Again, there is reason for pessimism. One obstacle is Mikkola’s point about the difficulties of convincing social agents to adopt the concept of woman-as-class and the question of whether it is worth investing scarce feminist resources in this project. If anything, the present proposal exacerbates this difficulty as the mission of convincing social agents not only to adopt the concept of woman-as-class but to do so only in certain contexts and in others to adopt the concept of woman-as-identity will if anything be an even harder one to accomplish.

Another obstacle is that the proposal is that it is unlikely to be successful in that there is reason to think that it won’t lead to a displacement of the old concept of woman. Consider the concepts of jade, jadeite and nephrite. These are different concepts. Initially, we had only one, i.e. the concept of jade. However, we found out that there are two different minerals, jadeite and nephrite, leading us to introduce the corresponding concepts. While the concepts of jadeite and nephrite are in use, the concept of jade remains in use also. This is because it is difficult to tell the difference between members of the extension of the concept of jadeite and nephrite and because that difference is not relevant for everyday purposes, at least for the most part. The worry is that the same holds for the old concept of woman, the concept of woman-as-class and the concept of woman-as-identity. After all, it is difficult to tell the difference between members of the extension of the concept of woman-as-class and the concept of woman-as-identity and that difference is not relevant for everyday purposes, at least for the most part.

5 Accordingly, we take examples like this one to provide compelling evidence against anyone who might claim that in order to possess a concept that admits of definition one must know this definition.

6 See also (Mikkola 2016, ch. #5).

7 Many thanks to an anonymous referee for pressing us on this point.

8 Crucially: We are open to alternative concrete proposals for what the linguistic innovation might look like. This is because we believe that this best not done unilaterally. Rather, it is most likely to be successful if the linguistic innovation is be done by a group of people with representatives from across all cultures. This is to maximise the probability that everyone can be on board with this and that no one is side-lined.

We would also like to emphasise that we are not even wedded to using a term that is associated with women. The reason we chose ‘she’ is that it follows the example of the precedent for the racial terms ‘negro’ and ‘black’ that we will discuss in more detail below. That said, we are open to the idea that an entirely new term may be the best option at the end of the day.

9 We are very grateful to an anonymous reviewer for pressing us on this.

10 One question one may have about our proposal is whether the case of ‘negro’ (and ‘black’) provides the best model for the kind of linguistic innovations that we should be aiming for. More specifically, one might wonder whether the case of ‘eskimo’ (and ‘inuit’) isn’t a better model. The key difference here is that in the case of ‘negro’, the social norm prohibiting use of the term does not extend to black people, at least not typically, whereas in the case of ‘eskimo’, it does.

We agree that this is indeed an important difference between those linguistic practices. And, of course, it may be that one of the two options is better at the end of the day. If so, there is reason to prefer one option over the other. That said, we won’t aim to settle this question here. Fortunately, we don’t need to. The reason for this is that the kind of moral progress that, as we will argue momentarily, our proposal brings into view can be made on either model. For present purposes, we’ll rest content with this result.

11 Note that this may happen independently of whether the assertion featured the term ‘negro’. Crucially, linguistic innovation cannot plausibly be expected to do all the work in eradicating reactions to racist generics. The claim we make is more modest: that replacing the term has brought forth different political connotations associated with the term. In this, what we are claiming is merely an improvement: rather than triggering racist reactions via two linguistic tools combined – i.e. the generic, and the racist term – replacing the racist term via a reclaimed one makes moral progress.

12 Doesn’t our proposal gloss over important disanalogies in the subordination and oppression profiles between the cases of race and gender? We’d like to thank a reviewer for this journal for raising this question.

We do not mean to deny that there are important differences here. Crucially, however, there is reason to think whatever differences there may be, they won’t stand in the way of our proposal contributing to moral progress in the gender case. After all, the way in which linguistic innovation promises to bring about progress in the case of race does not turn on specifics of the subordination and oppression profile. Rather, what it turns on are general social and conversational mechanisms of the sorts described above. As a result, there is reason to think that whatever disanalogies there may be between the cases of race and gender, they will not cause trouble for our proposal for linguistic innovation.

13 One further reason for this is that the proposal can avail itself of the resources provided by the proposals that have been already been made in the literature to ensure trans inclusivity. To take just one example, it is entirely compatible with our proposal to countenance a concept of -woman-as-identity, alongside the concept expressed by ‘she’ and to place the two concepts on equal footing. If this isn’t immediately obvious, note that we are not aiming to abolish women. Rather our aim is to abolish the use of the term ‘woman’ through linguistic and normative innovation. Introducing Jenkins’s concept of woman-as-identity (or she-as-identity as we would now prefer) is compatible with this so long as we don’t use the word ‘woman’ to express the concept or in its definition, for that matter. Crucially, we believe this won’t be necessary, since the concept associated with the term ‘she’ is trans-inclusive. That being said, even if we are right about the concept, it may be beneficial to explicitly employ Jenkins’s definition as a way to engineer mistaken, trans-exclusive conceptions about women.

14 See Simion (2017, 2018, 2019) for epistemic limitations to projects in conceptual engineering. See also Kelp and Simion (2021, 2017) and Simion (2021) for knowledge transmission and the normativity of speech.

15 It also promises to allow us to seize the opportunity that Jenkins would miss out on were she simply to abandon Haslanger’s concept of woman-as-class.

16 In philosophy of action intentional not doings of agents are often referred to as negative acts (e.g., Vermazen 1985).

17 One might wonder whether the concept that ‘woman’ comes to express in this scenario isn’t the concept of woman-as-class. After slurs have everything to do with subordinating their targets in virtue of certain salient features. On semantic accounts of slurs, these properties will be written right into the meaning of slur words. As a result, ‘woman’ will have a subordination condition in virtue of certain salient features written right into its meaning. And isn’t that just the concept of woman-as-class?

No. There are important differences between the concept that ‘woman’ comes to express in this scenario and the concept of woman-as-class. Most importantly for present purposes, it is part of the meaning of slur words, on semantic views, that the relevant forms of subordination ought to happen. In contrast, the concept of woman-as-class is either neutral or clear on the unjust character of the subordination. The concept ‘woman’ comes to express in this scenario is thus importantly different from the concept of woman-as-class.