Ethics Discussions at PEA Soup: Erich Hatala Matthes’ “History, Value, and Irreplaceability,” with précis by Carolyn Korsmeyer

27 Replies to “Ethics Discussions at PEA Soup: Erich Hatala Matthes’ “History, Value, and Irreplaceability,” with précis by Carolyn Korsmeyer”

1. I’d like to thank Kate Manne and Hille Paakkunainen of PEA Soup for inviting me to contribute to this discussion. I found Matthes’ essay to be both extremely interesting and important. I thought the criticisms of Cohen’s position were spot on, and the final section of the paper seems to me to be breaking important new ground for thinking about historical value and the value of history. I do have some concerns about the core argument. Very briefly, my worry is that Matthes suggests that he is defending a novel and surprising thesis regarding the separation of historical value from irreplaceability, but I believe the degree of separation he argues for ends up appearing plausible in large part because he adopts a broader notion of irreplaceability than is usually employed.
   (Though I’m coming at things from a slightly different direction than Carolyn Korsmeyer, I think it will be clear from what I say below that I’m quite sympathetic to her remarks about the difficulty of expunging irreplaceability from our thoughts about historical value.)
   Matthes’ thesis regarding separation appears in various forms. Here’s one version from p.40:
   “There are nonhistorical ways that an object can satisfy IR [irreplaceability] without acquiring the special character of objects that we value for their histories, and historically significant objects do not always satisfy IR in the way we might expect them to: hence irreplaceability can be neither necessary nor sufficient for securing historical value.”
   This can sound quite surprising at first, given that many discussions of irreplaceability have taken historical properties to be doing essential work in helping us understand the phenomenon though their capacity for individuating the relevant object (be it a loved one, artwork, historical artifact, etc.). A tight connection between irreplaceability and history is usually presumed, even if it has not always been explicitly explored by those previously writing on the topic.
   Matthes’ separation thesis turns out to be far from controversial, however, because the notion of irreplaceability that he adopts for the purposes of the paper is very broad. At first it is hard to see this, since the phrasing is fairly vague:
   “Irreplaceability (IR): An object is meaningfully irreplaceable if and only if all candidate substitutes would fail to be valuable in the same way as the original.” (p. 38)
   How broad this conception actually is comes out later, perhaps most explicitly on pp. 59-60:
   “Return to our trusty, instrumentally valuable umbrella. If there were only one umbrella left in the world and no more could be produced, that umbrella would be an exemplar of irreplaceability: it would be the only thing valuable in precisely the same instrumental way that umbrellas are. It satisfies MR, assuming that the instrumental value of the umbrella is indeed our maximally differentiable assessment of its value, and, because no suitable substitute can exist, it would also satisfy IR.”
   Now, if that is the sort of irreplaceability under consideration, then it is indeed quite obvious that “there are non historical ways that an object can satisfy IR without acquiring the special character of objects that we value for their histories.” The umbrella is simply instrumentally valuable and its value is completely ahistorical. It is also irreplaceable only in the sense that, as it happens, no others are around or available.
   However, I’m not sure that a conception of irreplaceability this broad is really the correct one to focus on given that the essay begins with a discussion of the way in which irreplaceability manifests in our thoughts about personal relationships, artworks, and valued portions of the natural environment. In particular, I’m worried that Matthes’ broad characterization unhelpfully conflates a relevant (narrower) notion of irreplaceability with what we might call “matter of fact” rareness or uniqueness.
   That the irreplaceability of the beloved (for example) ought not to be thought of in terms of mere rareness (i.e., the fact that no replacement is currently available) comes out when we consider that upon being convinced that a perfect duplication of the loved one would be possible it still seems entirely reasonable for one’s attitude regarding the irreplaceability of the beloved to remain unchanged.
   A passage from Harry Frankfurt nicely makes this point:
   “Imagine that one day a young woman turns up of whom I discover that she is, and always has been, indistinguishable in every discernable physical, psychical, and behavioral respect from one of my beloved daughters. I would find that bewildering, and it would certainly distress and inhibit me in various ways. But however confusing and disruptive the circumstances might be, they would surely not lead me to conclude that I had all along been somehow wrong to love my daughter because I had erroneously supposed that there was no one quite like her.” [“On Caring”, p.169]
   David Cockburn has expressed a similar view when discussing the qualitative or empirical notion of uniqueness that Martha Nussbaum appears to endorse in some of her writings on The Symposium:
   “There is a sense of ‘individuality’ which has been completely lost as soon as there is room for talk of the chances of finding a substitute; and the remoteness of the chances does nothing to change that. [….] If we understand the notion of irreplaceability in this way my thought of my son as an irreplaceable individual is a hostage to fortune.” [Other Human Beings, p. 151]
   So, my concern is that any conception of irreplaceability broad enough to count the merely instrumentally valuable umbrella as irreplaceable is not adequately focusing in on the phenomenon that has typically interested philosophers when they talk about a loved one (or work of art, etc.) being truly irreplaceable. My suspicion is that the relevant narrower notion of irreplaceability (which comes out in the quotations from Frankfurt and Cockburn above, but which can also be brought out through the various “duplicate swap” thought experiments popular in the literature) does in fact have a tighter connection to historical properties and to historical value than one might expect given Matthes’ discussion. (Though I think the situation is complicated, which shouldn’t be surprising given my discussion of the Hendrix guitars that Matthes cites.)
   There’s lots more to say, but I think I’ve probably already taken up too much space for a blog post, so I’ll end for now. I do want to reiterate that even if my particular concern here turns out to be justified, I think Matthes’ essay contains a wealth of valuable insights, and I want to thank him for taking the time to think so carefully (and fruitfully) on these topics.

2. First off, I would like to thank the folks at PEA Soup for hosting this discussion, in particular Kate Manne and Hille Paakkunainen, who were the primary organizers. I am deeply grateful for this opportunity. Second, many thanks to Carolyn Korsmeyer for her generous and thoughtful précis, which I will turn to now. I want to begin by clarifying a matter that Korsmeyer calls attention to, and I hope that this will aid me in responding to the distinction that she proffers between “an object that has historical value and object that is valued for its history.”
   In explaining part of my view, Korsmeyer writes: “Properties deemed ‘historical’ are normative features that only things of special and lasting importance possess. Anyone might cherish a keepsake for its singular history, regard it as irreplaceable as such, and at the same time grant that it has no historical value at all.” I’m not sure that I actually want to accept these claims, though if I appear to in the paper, that is no doubt due to a lack of clarity on my part. This paper stems from a larger project (my dissertation), and I may have neglected to sufficiently establish ideas in this piece that are expounded in greater detail elsewhere. So, as it turns out, I did not intend for the “historical” in “historical value” to refer only to the kind of history with which, for instance, historians are concerned. Indeed, I think the scope of historical value can vary widely from the personal to the universal (or at least approaching the universal), and so did not mean to use the term in the historian’s sense. Rather, I take “historical value” to refer to any object that is appropriately valued for its historical properties, understood as properties pertaining to its past. This is what I was attempting to establish in the note on terminology in footnote 2. Accordingly, on my view, a person who cherishes a keepsake for its singular history, and believes it appropriate to do so, should be prepared to defend that keepsake’s historical value, even if it is historical value of a very personal kind. Granted, a person might find herself with positive evaluative attitudes towards an object and simultaneously not grant that it has historical value, but this would be tantamount to acknowledging that her evaluative attitudes are unwarranted.
   Another way of getting at this point is to say that I am not nearly as concerned about the proliferation of historical value as I am about the proliferation of irreplaceable value. This is why when I first introduce the proliferation problem (39) I cast it in terms of the “unacceptable implication that everything is meaningfully irreplaceable,” and I believe (hope!) that I continue to confine my remarks about the proliferation problem to irreplaceability throughout the paper. Indeed, I think there is a lot of historical value in the world, value that comes in many kinds, degrees, and scopes. This is why I believe that historical value provides such an interesting topic for philosophical reflection, and why it is so important to articulate the independence of the rich concept of historical value from the bluntness of irreplaceability.
   My hope is that I can now use this clarification to explain what is going on in Korsmeyer’s example of the mourning brooch. To the collector, various mourning brooches may be replaceable one for the other because she has no basis for distinguishing among them with respect to their historical value. On the other hand, to a specific widow, a particular brooch is irreplaceable because it contains her husband’s hair. (This kind of difference in value is unsurprising given the relational nature of historical features. While the potential for historical value in a given object will be delimited by the facts of its past, its value can differ for different people based on the relation in which they stand to its history.) I’m not sure I completely agree with Korsmeyer’s assessment of what makes the brooch irreplaceable to the widow. She says that this is because the widow’s fallen husband “and only he was the nonfungible intentional object of her emotional attachment, and nonfungibility transfers to the token of his hair.” However, while I can agree that the value of the brooch is a function of a transference of value from the husband (if one wants to describe it that way), I do not think that nonfungibility transfers with it. If what she ultimately values about the brooch is that it contains (literally) a piece of her husband, then on my view, she has just as much reason to value any other brooch that contains her husband’s hair. If the brooch is irreplaceable, it is irreplaceable not just because her husband is, but because it is the only brooch with the right historical features to justify her particular evaluative attitudes. If we can say that it would be a certain kind of mistake for her to value a brooch containing someone else’s hair (as I think we should), then we should also say it is a mistake for her to regard one brooch as irreplaceable if it has the same evaluatively relevant features as another. As I try to argue in the paper, we avoid this conclusion only by giving up on a relationship between our evaluative attitudes and reasoned justification, or by giving up on the idea that reasons generalize, both of which have far more troubling consequences than the idea that irreplaceability turns out to be contingent (I will discuss the contingency issue further in response to Chris Grau’s comment). So I think we can explain the potential irreplaceability of the widow’s brooch without opening up the proliferation problem after all.
   I am wary of going on for too long, so I will wrap up for now, though I of course encourage everyone to press me on further aspects of Korsmeyer’s discussion that I did not adequately address, and I want to thank her again for her comments. But first, I think it is worth mentioning that I am unmoved by the psychologist’s charge of “magical thinking” (though I thank Korsmeyer for drawing this literature to my attention, and I look forward to checking it out). To my ear, “magical” suggests something that is either false or defies explanation, but neither seems to be true in the case of historical value. After all, I have argued that what is so captivating about historically valuable objects is that they really did occupy a different time: indeed, it is that the historical descriptions of certain objects are true that makes sense of our evaluative attitudes. As such, I do not think talk of reasons and justification is out of place in this domain.

3. Thanks to Chris Grau for these excellent comments. Grau suggests that the notion of irreplaceability that I employ in the paper “unhelpfully conflates a relevant (narrower) notion of irreplaceability with what we might call ‘matter of fact’ rareness or uniqueness.” He illustrates this narrower notion of irreplaceability with respect to the irreplaceability of one’s beloved. Now, I do discuss the case of persons throughout my paper, and I appeal to some of the philosophical literature on love to make some helpful distinctions. However, I did try and avoid using personal relationships as one of my central cases. This is because it is the domain in which people seem most compelled to give up on trying to justify or account for their evaluative attitudes in terms of reasons that can generalize, in favor of a view on which the beloved is valued just for being the particular person that he is. I discuss in the paper (largely building on the work of others) why I think that’s a dangerous move (46-47, in particular). However, given that he thinks that this narrower notion of irreplaceability “does in fact have a tighter connection to historical properties and to historical value than one might expect given Matthes’s discussion,” I take it that Grau is comfortable making reference to the historical features of the beloved and one’s relationship with him in order to explain his irreplaceability.
   In that case, I think that any justifiable account of irreplaceability must ultimately be the kind of contingent “matter of fact” account that I argue for, even in the case of personal relationships. In the case of the beloved that is often invoked here, it will just turn out that a replacement is impossible in the relevant sense. Consistent with my argument and the rest of my examples in the paper, it all comes down to the features that can justify one’s evaluative attitudes. So, to take the Frankfurt case, let’s suppose that the way he values his daughter is based in part on the particular history of valuable interactions between them. It’s not clear from Frankfurt’s example whether this doppelganger who turns up is supposed to also share that history of valuable interactions with Frankfurt (N.B. Word wanted me to change the “who” after “doppelganger” to a “that”: doppelganger’s are people too!). If not, no problem! Doppel-daughter lacks important historical features that explain Frankfurt’s evaluative attachment to his actual daughter: turns out she’s irreplaceable to him, just like he thought. However, were we to imagine, per impossibile, that his daughter suddenly “twinned” such that there were now two of her, and each must be thought to share the same history with him if either does, then surely he would have reason to love and value them both: what possible basis would there be for differentiating them, evaluatively or otherwise? But of course, this is all crazy talk. It just doesn’t matter that irreplaceability is contingent in this context, because in the case of the beloved daughter, Frankfurt will never in fact confront an evaluative substitute.
   I think we can be led astray by focusing too much on the case of beloved relationships. While the possibility of an evaluative substitute in such cases may only be metaphysical, evaluative substitutes in many other cases of historical value can exist in this world. How often they do will depend on how fine-grained we can get in explaining the relevance of an historical feature to an object’s historical value. But as long as justifying the claim that an object is valuable in a certain way involves citing a feature of that object (historical or not), it remains open that something else might be valuable in that same way, too. Thus, I think the contingency of irreplaceability is something that we have to accept if we want to be able to pursue such a justificatory strategy.
   Now, there may be further reasons for thinking that one ought not be open to entertaining evaluative replacements, or that one ought not seek them out. However, that is different from the claim that no evaluative replacements exist. I don’t think we necessarily need to think the latter is true in order to believe that the former may be appropriate in certain contexts. Admittedly, though, I find it hard to see what context such a resistance to entertaining evaluative replacements might arise in except for cases of loving relationships. Moreover, I think such cases are the ones in which evaluative replacements are, as it happens, least likely to exist.

4. I don’t see how the account of what makes historical valuing objects such as the pen that was used to sign the emancipation proclamation appropriate can work for items of mere personal significance.
   The account holds that it is appropriate to historically value objects that have a significant history — “a significant connection with a significant past” (p.30). From what I can tell, the account is a little stronger than it might at first appear. By appropriate, I think it means that we have pro tanto reasons to value these things, not just that it is permissible. (The discussion of Moller’s paper on 22-23 suggests this reading.)
   My worry is that the key condition just doesn’t apply to objects of sentimental value, such as mementos and childhood haunts. They aren’t significant. However, if it does apply, then it results in something much worse than the proliferation problem. It suggests that I have reason to value your objects of sentimental value.
   Note, on this account the appropriateness is based on the value of the object. The significance is attributed to the object, not the valuer. It’s not that I find it significant that justifies my valuing; rather, it’s that the object is significant (or the history to which it is connected is significant). Hence, if your grandma’s tea cup has significance, then it has significance. The problem is that I don’t have any reason to value your grandmother’s tea cup (other than that you would be upset if it were destroyed).
   I don’t see anything in the paper that could explain why some objects are significant to me and not to you. The only way I see out of this is to say that mementos have agent-relative significance. But I sure wouldn’t want to go there. I can’t make a lick of sense out of the idea.

5. Hi Aaron, thank for your comment. To say that something is valuable, as I understand it, is (at least) tantamount to saying that someone has reason to value it, or that it is appropriate to value it. However, I don’t think we need say that if something is valuable then everyone has reason to value it. Rather, the scope of values can vary. So I agree, you don’t have reason to value my grandmother’s tea cup, because you lack the relationship to my grandmother that would make it’s history have evaluative relevance for you. But I don’t think that should lead us to give up on claiming that my grandmother’s teacup is valuable. Indeed, if I’m about to chuck the teacup, I think my sister might remind me that I have reason to value it (given some assumptions about the teacup and my grandmother), and that reminder would have purchase with me. If we analyze sentimental value in terms of mere desiring or liking, then it seems to me we give up on the ability to talk about the appropriateness of our attitudes (or lack thereof) towards the objects in question.
   This is a way of looking at the issue starting on the very personal side, but you can also approach it from the more universal side. What exactly is it that makes valuing the pen that was used to sign the emancipation proclamation appropriate? Do you think it would be appropriate for anyone to value it? Why? Why think there is such a sharp divide between the things that are intuitively thought of as appropriate for anyone to value (things of “world-historical” significance) and things that is appropriate for only one or a few people to value? I think there is much to be gained from thinking of these issues on a continuum.
   I realize that this is contentious, and I try and defend these views at greater length elsewhere. Hopefully it at least provides a sense of where I’m coming from in making related claims in the paper. Alas, space is always limiting in print, and the paper was already quite long.

6. Thanks for the reply. I’m still not entirely clear about the account on the table. Are you endorsing a buck passing account of value? I thought the account was that the historical properties give the object value and the value is what gives us reason to value it.

7. Hi Erich,
   Thanks for your reply — it really helps to clarify your position. By my lights you are really biting the bullet here in denying that there’s a relevant sense of irreplaceable that does not amount to mere “matter of fact” rareness. It seems to me David Cockburn is saying something true and important in the passage I cited earlier. I also think that giving up this notion of irreplaceability because of justificatory worries is wrongheaded: if something has to go here, it is an overly rationalistic sense of what sorts of justifications are possible in this context. (As you know, this is the sort of approach I argue for in “Love and History”.)
   Regarding your example of the beloved dividing. When discussing objects dividing you say:
   “You have no possible basis for differentiating their values, and they are therefore substitutable for each other….”
   I take it given your remarks above this is the same view you have of the two people after the fission. But it seems to me there is a perfectly coherent and important sense in which those two individuals are indeed individually irreplaceable, and the thought of treating them as substitutable with each other is rather problematic. (Though of course I also agree that “surely [one] would have reason to love and value them both”.) I have similar (though less strong) intuitions about cases involving artworks and some other objects, but I may be eccentric in that regard. I look forward to hearing the views of other folks on these issues.

8. Aaron, while I have some buck-passing sympathies, I don’t think that the present discussion hinges on questions of normative vs. evaluative priority. All that need be the case is that values need not necessarily issue in reasons for everyone to value them. I think it is usually most helpful to think about values primarily in terms of how and for whom it is appropriate to value them, and perhaps that perspective is best accommodated by a buck-passing account, but it’s not clear to me that it requires or entails it. Indeed, some have critiqued the buck-passing account for not being able to account for the true diversity of values.

9. I’m just trying to figure out what exactly accounts for the difference between whatever reasons you might have to value your grandmother’s teacup and my lack of reasons to value it. It’s connected to a significant past, or so we can assume. So you want to say it is valuable. Fine. The difference in what reasons you and I have to value that valuable object seem to stem from the fact that past which gives it value is yours and not mine. But why does that make a difference? Intuitively it does make a difference, but why? I just don’t see it accounted for here.
   We need an example. How about Butch’s watch from “Pulp Fiction”?

   Or, less offensive, the coin from “No Country for Old Men.” Don’t put it in your pocket:

10. Hi Chris,
    I guess I don’t really feel like I’m biting any serious bullets, but I think that’s because I really don’t feel the force of the person-fission cases. For one thing, we’re talking about people here, so once the fission occurs, they would quickly develop separate histories of the kind that could ground a differentiation in value with respect to the relationship. It’s very important to keep in mind that we’re talking about their value in that context: I’m not endorsing a view on which, for instance, the destruction of one person might be made up for by the creation of another from some impersonal perspective. But second, even if we screen off the individual trajectories that their lives will take and focus on the very moment of fission, I don’t think there’s anything problematic about treating them as substitutable in an evaluative sense. After all, all I mean by that is that the two are valuable in the same way. Our intuitions about the uniqueness of persons can make that seem weird, but in a world of person-fissions, I doubt we would have the same intuitions about individuality and uniqueness that we do in this world.
    Moreover, once we move away from the context of persons, the refusal to consider the possibility of evaluative replacements can even seem obstinate. If you value an object for a certain set of historical features, and I provide you with another object that shares those same features, I’m not sure what the evaluative merit would be in your insisting that you are unwilling to consider the possibility that the two are valuable in the same way. What would that demonstrate about your evaluative attitudes that we would want to hang onto in our account? To my mind, it would just demonstrate a kind of close-mindedness, and a failure to respond appropriately to the consequences of one’s own evaluative commitments.

11. Hi Aaron,
    Well, to be fair to me, I don’t think I claimed to be offering an account of that phenomenon in the paper, though I do of course make use of it. Why are you dissatisfied with the idea that the significant past that makes Butch’s watch valuable might be significant to him but not to Marcellus Wallace? Historical significance is plausibly relative to a context of other assessments of significance and value. I doubt there is anything it means for something to be historically significant full stop. So why not think that that the relative aspect of the historical significance of the watch is transferred to its historical value; that is, to for whom it would be appropriate to have evaluative attitudes towards it on the basis of its history?

12. Hi Erich,
    You say: “Our intuitions about the uniqueness of persons can make that seem weird, but in a world of person-fissions, I doubt we would have the same intuitions about individuality and uniqueness that we do in this world.”
    I agree that in a world of routine “person fissions” we’d have different intuitions, but that doesn’t change the fact that in our world the idea that persons (and other things) are irreplaceable (in the sense I’ve argued for) is an important one for many of us.
    I’m going to shut up and let others talk, but briefly: I suspect we are not going to convince each other on this issue. (The difference goes deep and involves perhaps fundamentally different philosophical sensibilities.) Regardless, your account includes a bounty of valuable insights that don’t hinge on this difference between us, and I’m hoping it will spur more writings on this still neglected topic.

13. Thanks for your kinds words, Chris, and for your questions, which I will continue to ponder!

14. I too thank Hille and Kate for inviting me to participate in this discussion.
    I’ve enjoyed thinking about the ideas in Erich’s dissertation since I first encountered them more than a year ago.
    I want to continue to press Erich on whether there is another notion of irreplaceability that might be more closely tied to historical significance. When Erich says that the last hammer in the world (when it can’t as a matter of fact be replaced) is “irreplaceable” it seems that he’s using a very different notion of irreplaceability than one that intuitively seems tied to historical significance.
    Erich mentions and, if I am understanding properly, sets aside Cohen’s remarks about an object’s being *destroyed* and then replaced. But perhaps destruction is at the heart of the matter. What about this notion of irreplaceability?
    X is irreplaceable iff if X were destroyed and replaced with a duplicate (even a duplicate with similar relational and historical properties) then a serious loss would have occurred.
    When my beloved new party dress is stained with red wine and I buy a new one that is an intrinsic duplicate, no serious loss has occurred.
    When one’s beloved spouse dies and then one quickly remarries, Erich finds this disturbing. (p 22) After all, a serious loss has occurred. But the quick remarriage need not involve any failure to feel a serious loss; it is simply one path among may *through* that loss.

15. Thanks for your comments, Liz! I have at times been attracted to analyzing irreplaceability in terms of destruction. I agree that the feeling of irrevocable loss that can accompany something’s destruction can seem indicative of its irreplaceability. My worry about this approach is that I think it may conflate two separate issues.
    So consider the case that I appeal to when I first try and put aside matters of destruction: “if we assume that two Warhol silk screens from the same series are valuable in precisely the same way, then, other things being equal, it would follow that you ought to accept one as a substitute for the other, and neither is strictly speaking irreplaceable. But this is compatible with it being a very bad thing if one of the Warhols were thrown on the bonfire” (38).
    If one of the Warhols burns, that is a serious loss in one sense: namely, something of great value has been destroyed. But in a different sense it’s not a loss, because there is still something in the world that is valuable in the same way as the artwork that was destroyed. Irreplaceability, as I have analyzed it, pertains to this second understanding. Because the destruction analysis and my analysis do not rise and fall together, I think it’s important to keep them separate. Indeed, as I understand it, it would be a mistake to infer that because the destruction of something is a serious loss, there must therefore not be anything else that is valuable in that same way (as in the Warhol case). Moreover, it may be that the loss of something uniquely valuable does not constitute a serious loss: while it may have been valuable in a different way from other things, it may not have been all that valuable in the first place.
    Importantly for my purposes, I think it is my sense of irreplaceability that people often invoke when they claim that historically valuable objects in particular are irreplaceable: they think that because histories are unique, each thing valued for its history must be uniquely valuable. I try to demonstrate in the paper that such an inference can also be mistaken. So, it may be that my account of irreplaceability could be usefully supplemented by a more explicit account of the serious loss of destruction. But I don’t think such an account could replace the one that I have offered in the paper.

16. Thank you, Erich!
    Consider this claim:
    (*) If something has historical significance, then it is irreplaceable.
    You deny claim (*).
    My thought is that there is a sense of “irreplaceable” such that X is irreplaceable iff if X were destroyed and replaced with a duplicate (even a duplicate with similar relational and historical properties) then a serious loss would have occurred.
    And on this sense of “irreplaceable”, claim (*) is true.
    As I understand it, your view is that the case of the two Warhol screens shows claim (*) to be false:
    You claim that each Warhol screen is obviously replaceable, and yet each has historical significance.
    You’re right of course that in some sense of “replaceable”, each Warhol screen is replaceable with the other. But I’m not convinced that the Warhol screens are replaceable in the sense I have proposed.

17. Thanks, Liz, that’s helpful.
    So it looks like we agree that some X can have an evaluative replacement in my sense, but nevertheless be irreplaceable in your sense: you agree that in my sense the Warhols are replaceable, one with the other, and I took myself to be agreeing that they are irreplaceable in your sense, because the destruction of either, even granted the availability of the evaluative duplicate, would be a serious loss.
    Looking at your sense of irreplaceability, though, I wonder if it sets the bar too low, and thus is subject to a proliferation problem parallel to the one I discuss in the paper. For one thing, the notion of a “serious loss” is doing a lot of work in your sense of irreplaceability, and it would be helpful if we could spell out what that amounts to. Because at first blush, serious loss looks like it might not actually be so serious on your account. After all, you say that on your sense of irreplaceability, claim (*) is true. But that seems to suggest that it is sufficient for serious loss to have occurred that the X in question be historically significant. Based on the capacious sense of historical significance that I’ve been working with, that looks implausible. It imputes to historical significance a gravity that may not be applicable: I think it’s important to acknowledge that historical significance comes in degrees.
    Now, among your options would be to take issue with the broad sense of historical significance that I employ, or to raise the bar for serious loss such that the destruction of any historically significant X need not be a serious loss. Of course, if you take the latter approach, then claim (*) is no longer true, even on your sense of irreplaceability, and that’s all I need to secure my desired conclusion. If you want to take the former approach, then we would need to talk more about what the right notion of historical significance should be. There of course also may be further options for your account that I’m not thinking of here.

18. Is the general argument something like this crude formalization?
    1. Our valuing mementos is often rationally justifiable.
    2. In order for our valuing mementos to be justified, it must be based on the historical properties of the memento.
    3. Often, many different objects can have the same historical properties.
    4. Hence, if the historical properties are what justify our valuing mementos, then when more than one object has the same historical properties, we have no more or less reason to value any one more than the others.
    . . .
    5. Hence, mementos are only irreplaceable if they happen to be matter of fact unique.
    It seems like you started with an assumption that our valuing irreplaceable objects is rationally justifiable and then tried to figure out how. You landed at historical properties. And then this lead you conclude that irreplaceability is only matter of fact uniqueness.
    I’m not convinced that we should start by assuming that the attitude is indeed rationally justifiable. I guess my worry is that we risk misdescribing the attitude if we start by assuming that it can be justified. It seems that the attitude we have toward irreplaceable objects is one that doesn’t accept substitutes with historically identical properties. Hence, I don’t buy that we value the historical properties in most of the clear cases. Grau seems right that we value the individual who is picked out by the history, not the history, or at least not primarily. This is very clear when it comes to love of persons and, even, pets. I might value our history together. Sure. But I am attached to and I care about the person. It seems that something similar is the case when it comes to irreplaceable objects.
    Perhaps this can’t be justified. But I see no reason to worry about this. Why think that it’s a mere bias? Why think that it is irrational? Perhaps that’s just how we are and there is ultimately nothing really wrong with it. We aren’t Vulcans. . . .
    On another topic: In reading your paper I kept waiting for you to talk about photography. The idea that we value mementos for the way in which they help connect us with the past has strong echoes of Bazin (and Barthes and Walton) on photographs. It would be interesting to sort out the differences between the way in which we cherish photos and other mementos. It seems that photos provide a strong connection, but we often value trinkets more than pictures.

19. Hi Aaron,
    Yes, that’s roughly the shape of the argument, and I am definitely starting with an assumption (or at least a hope!) that our evaluative attitudes are justifiable. Though I don’t just happen to land at historical properties in thinking about what justifies our evaluative attitudes towards certain objects. Rather, many (though not all) who talk about valuing objects for their histories already appeal to the historical properties of the object. Indeed, it is because they cite such properties that I take it we are already in the domain of justifying our evaluative attitudes. This is particularly so when we move away from the focus on loving persons, and I have already lodged some hesitance about framing the entire discussions in those terms. It would be strange, for instance, for such a paradigm to dictate the terms of how we value all archaeological objects, a different kind of exemplar of historical value.
    If you’re happy to deny that assumption and give up on the idea that evaluative attitudes are justifiable, then my argument probably won’t be of much interest to you. However, I think we give up quite a bit when we make that move. You think “that we risk misdescribing the attitude if we start by assuming that it can be justified,” but I think we risk failing to explain important aspects of our evaluative behavior if don’t. If our evaluative attitudes lack justification, it’s not clear to me how we can criticize or commend their appropriateness in the way that we often do (and I think ought to do). But, as Chris noted, this seems like a rather fundamental difference in approach that we may not be about to reconcile here.
    I should note, though, that I don’t think valuing some X for its historical properties should be taken to imply that it’s only the historical properties, and not X itself, that we value. As I discuss in the paper in conjunction with some of the literature on love, this confuses the basis for valuing and the focus of the evaluative attitudes. Some philosophers have insisted that the property-based reason that is cited for valuing X must therefore be the thing that is valued, but, to my mind, this has been no more than an insistence. The distinction between the basis and focus of valuation explains why that concern is unwarranted, and I have yet to see a compelling objection to that distinction.
    Your comment about photography is very interesting, and I admit that I have not thought about it much yet. It’s definitely a point I should consider further. Thanks!

20. Hi Erich,
    It’s been fun to read the paper and all of the comments.
    I was interested in your discussion of the Kraut-blanket type cases and the notion of substitution classes. Do you have thoughts on how these get fixed? Presumably the same object can have a wider or narrower substitution class for different people, depending on how they relate to it. And those relations, I take it, can be pretty willy-nilly. Whether I latch on to something in virtue of some experience I had with it need not be a response to reasons that would command respect from others similarly positioned, and yet those responses could determine that the substitution class is narrow for me but wide for you. If so, is whether something is irreplaceable ultimately not just contingent but subject to fairly arbitrary vicissitudes in our responses? I guess that would encourage a blase, deflationary attitude toward the irreplaceable.

21. Hi Dan,
    Thanks for your question. Yes, the same object can definitely have a wider or narrower substitution class for different people, depending on the different ways in which that object might be valuable to different people. I should note that though this can occur in particularly stark ways in historical cases, I think it’s a ubiquitous evaluative phenomenon. I may value our weekly pickup soccer game for the exercise and because it gets me outdoors, whereas you may value it because of elements of soccer in particular (its specific skill-set, its singular type of beauty, etc.). I would be just as happy to play something else; you might only show up because we’re playing soccer. We value the same activity, but we value it in different ways, and that sets different substitution classes for each of us.
    Now, because I think that our evaluative attitudes can be appropriate or not, the substitution class for a given valuable X will not just be set by whatever attitudes I happen to have towards it: rather, it will be set by a combination of the features of X, and the attitudes it would be appropriate for me to have towards X given those features (which may also range depending on features of me, such as my other evaluative attitudes). So while it’s true that I may or may not “latch onto something in virtue of some experience,” my thought is that whether or not it is appropriate for me to latch onto it will depend on the underlying set of features I mentioned above. So I disagree with part of your characterization: my latching onto that X would be a response to reasons that would make a similar attachment appropriate for anyone similarly positioned. It should be emphasized that appropriateness should not be analyzed in any kind of obligatory or compelling way: I’m not forced to value everything that I have a reason to value, though it may be that I am missing something if I don’t happen to value it. That’s not necessarily a criticism, though: we only have time for so much valuing in our lives.
    That’s just a sketch of course: I’m working on spelling this out in greater detail elsewhere. However, I don’t think this picture encourages a blasé or deflationary attitude toward the irreplaceable. Indeed, I think it helps us take a discerning approach towards assessing which things in our lives are in fact valuable in ways that no other thing is. Importantly for my purposes in the paper, it will allow us to realize that not everything we value for its history is irreplaceable.

22. First of all, I would like to thank everyone involved with PEA Soup for the invitation to participate in this forum, and also those who have posted these thoughtful comments. It is a great way to engage in philosophical debate that is quick, informative, and civil. And, of course, thanks to Erich for providing such an interesting paper for discussion.
    In his reply to my comment, Erich notes that he is more worried about the proliferation of objects deemed to be irreplaceable than with the proliferation of objects of historical value. Since I don’t see how the two values can be entirely separated, the problem of proliferation remains. But it helps to point out that the strong sense of irreplaceability refers to particular things rather than things of a type that still has a lot of tokens available. Things of historical value are often (though not always) rare, so when they are lost so is access to a remnant of history. This is just a note to add to the other posts that query irreplaceability in more detail.
    Although the lines separating them are not clean, I do think we need to distinguish old things for which we hold affection from things of historical significance. Both are cases of evaluative attachment, but do they have the same grounding? There are some divergences of intuition at work here, for Erich suggests that even an object of purely personal, sentimental value is cherished in part because of its historical value. Grandmother’s tea cup is a good example. But if the tea cup was mass produced and many of its tokens remain, then my attachment to one particular cup is surely a function of my attachment to Grandma, not a recognition of the place in history that her china possesses. Similarly with the mourning brooch, where the irreplaceability of the woven mat of hair is entailed by the nonfungibility of the love for the departed whose hair it was. (Naturally, it could be mounted in a different brooch, which is just a vessel for the hair, which is the thing that does not bear replacement.) Both cases, I still think, reveal the distinction between an object valued for the history it embodies (which might not in fact be “significant” history at all) and an object of historical value (a term that extends beyond sentimental attachment and connotes something of more general importance).
    Erich’s reply clarifies that what he means by historical value is not the sort of thing that historians might target. I am guessing that pursuing his line of thought might blur my distinction between objects of historical value and objects valued for their histories. This is an intriguing remark that invites further investigation of just what it is about the past that keeps its hold on the present.

23. Thanks for following up, Carolyn! Let me take this opportunity to say a bit more about why I think historical value must lie on a spectrum from the personal to the universal, and hence why I don’t think a distinction can be made between historical value and objects valued for their histories.
    In the paper, I refer to a passage from the late Arthur Danto’s Analytical Philosophy of History, where he writes: “a particular thing or occurrence acquires historical significance in virtue of its relations to some other thing or occurrence in which we happen to have some special interest, or to which we attach some importance, for whatever reason.” That’s a rather open-ended formulation of historical significance, but also, I think, an apt one. When we follow the trail of inquiry into why something is historically significant, and why it is appropriate to value it as such, we arrive at some other evaluative assessment, one that may be relevant to many people, or only a few. I see no non-arbitrary reason to limit this kind of account only to those things whose historical significance is relative to evaluative assessments of general importance.
    I think we can illustrate this point by moving through a series of historically significant places and things. Begin with a place that is thought to have a quite general historical significance, such as the battlefields at Gettysburg (I should add, in memory of Danto, that everyone should read his beautiful essay about Gettysburg in the Spring 1987 issue of Grand Street, which was first brought to my attention by one of Carolyn’s excellent papers). Now consider a place of somewhat more local interest, perhaps the site of a first state house that has since burned down, but is memorialized with a plaque or statue. Now consider a small town, where the library has a display case for the first book ever lent there. Now consider a family that makes an annual pilgrimage to the farm where the matriarch was raised. At what point along this progression do the evaluative attitudes towards objects and places of historical significance undergo a change in kind rather than merely one of scope, as I would put it? There are some big leaps here, so it’s not as if this is a Sorites problem: I just don’t even have a general sense of where we would make the distinction. There are specific historical features that justify the evaluative attitudes one might have towards Gettysburg, as well as those one might have towards a family farmstead. It is these kinds of considerations that lead me to think there is no non-arbitrary distinction between historical value and valuing something for its history, though I would be curious to hear more about how that distinction might be motivated in light of these concerns.

24. We learn a great deal from this thought-provoking paper about the value of things with special histories, and how irreplaceability bears, and doesn’t bear, on other familiar ideas: intrinsic value, unconditioned value, final value, historic value, and more. I want to think about some implications further afield.
    When something is fungible, it can be replaced by something else that can do the job just as well. ‘Doing the job just as well’ sounds like a description of instrumental value, and discussions in ethics and feminist philosophy have often tended to put ideas of fungibility and instrumentality together. Someone might think a one dollar bill is replaceable by four quarters, because they can do the same job, as means to achieving certain shopping ends. Someone might think Jane is replaceable by Julie, because they can do just the same job, as means to achieving certain sexual ends. The idea of replaceability has had an important role to play in Kantian ethics, and in feminist approaches to objectification. Kant’s claim that we are beings whose value is ‘unconditioned’ was thought to rule out treating each other merely as ‘means’ or instruments, or as replaceable. It underlies Kant’s distinction between ‘dignity’ and ‘price’: ‘What has a price can be replaced by something else as its equivalent; what on the other hand is above all price and therefore admits of no equivalent has a dignity’, as he puts it in the Groundwork. In Nussbaum’s work on objectification, ‘treating as fungible’ appears as one of the ways to treat a person as a thing, alongside instrumentality (unlike Kant, she does not assume that replaceability and instrumentality will coincide). It is a common theme in feminist work that objectification includes a refusal to acknowledge a person’s ‘dignity’ in Kant’s sense, treating women as if their value is their price, their market value.
    One initial lesson from Erich’s paper is that on the face of it, Kant’s supposed contrast plays out differently than he thought. Many ordinary objects are treated as if they have ‘dignity’ in Kant’s sense. Erich and his interlocutors agree on this: even the old eraser appears to have ‘dignity’ in this special sense. But Erich’s distinctive message is that irreplaceability is less important than we thought, and that Kant’s package of concepts can be unraveled. The world’s last umbrella is irreplaceable, but not intrinsically valuable. The mourning brooch is more than instrumentally valuable, but is replaceable—not of course by a mere intrinsic duplicate, but by a comparable brooch with the same historic connection to the dear departed. All this is compatible with our rightly valuing the umbrella as a mere means, and the latter as, perhaps an ‘end’ (depending how we interpret the remarks about ‘final value’).
    I want to suggest that there is one part of the package that Kant gets right, and Erich misses. The umbrella and the mourning brooch have more in common than Erich allows, and that this has implications for our taxonomy of value. Umbrella and brooch both have, in Kant’s terms, conditioned or extrinsic value (near enough—though I’ve elsewhere argued that even this is not quite right). The value of the umbrella depends on its useful effects. The value of the brooch depends on its historic connection to the departed loved one. In neither case is the value ‘intrinsic’ in the sense of Moore or his conceptual successors. In isolation from such relationships, neither would have any value. Instrumental value is one species of extrinsic value, depending on something’s possible effects in the future. Historic value is another, depending on its actual causes in the past. I want to suggest that the brooch’s supposed ‘final’ value is beside the point. In both cases we have an item with extrinsic or conditioned value—because of its usefulness, in one case, or its historic relations in the other. We value the umbrella as means, because it has instrumental value. We value the brooch not merely as a means, because it has more than instrumental value. But our language for this latter case is far from adequate.
    We have a word for valuing something because of its instrumental value: we say, valuing ‘as a means’. We don’t have a word for valuing something because of its intrinsic value. We’re tempted to say, ‘valuing as an end’, but where that signifies ‘not merely as a means’, it includes valuing for its conditioned, extrinsic value, as well as for intrinsic. That was part of Kant’s point: means-end valuing fails to match up with extrinsic-intrinsic value. Happiness, according to Kant, has conditioned, extrinsic value, being desert-dependent; but we value it as an end, not a means. Talk of ‘final value’ seems misleading, since it shifts an idea about the structure of our motivation to an idea about the structure of value. Back to the case at hand: we could say that the brooch is valued ‘as an end’, even that it ‘has final value’, provided we are clear that this is compatible with its having value that is still merely conditioned or extrinsic.
    The value of the umbrella and the brooch have, then, greater structural similarity than Erich allows. I have not addressed the question whether some relations, e.g. historic ones, are more evaluatively significant than others, e.g. instrumental ones. They may well be, and Erich does a wonderful job spelling out how certain ordinary objects can be eloquent time-travelling witnesses to the past. What though of his surprising claim that the mourning brooch, and not the umbrella, is replaceable? Here too I want to suggest that Kant has it right, and Erich has it wrong. Agreed: the mourning brooch is replaceable by another that fulfills the same historic conditions. But the umbrella too is replaceable. ‘Replaceable’ is a dispositional concept, to be unpacked in terms of the counterfactual: ‘could you replace it with another?’ To that the answer is, yes. A glass all on its own is fragile. And an umbrella all on its own is replaceable too. The umbrella could be replaced be another that ‘does the job just as well’—in instrumental terms; and whether or not such an object is actual. The mourning brooch could be replaced by something else that ‘does the job just as well’—not in instrumental terms, but historical; and again, whether or not such an object is actual. And perhaps Kant is closer to the truth than we allowed, in his initial contrast between people and things: human beings are not, in this way, replaceable.
    Getting back to people, Erich’s work provokes one further interesting question. Kant said we should not treat people merely as means. It is, he thinks, a moral mistake to treat people as having merely instrumental value. Would it be a moral mistake to treat people as having merely historic value? If what is bad is failure to acknowledge the ‘unconditioned’ value of human beings, then treating them as having any merely conditioned value should be bad. We perhaps value others, for example family members, in part for the specific history they share with us. Do we value certain others merely for their history? I don’t know. Perhaps an anthropologist might value members of a remote tribe in the way an environmentalist values the redwood trees in the forest Erich describes. But such pathologies, if they exist, are dwarfed by the pathologies of treating people as mere means to their ends.

25. Many thanks, Rae, for these wonderful comments. You call attention to an important ambiguity in the paper: apart from whether or not replaceability ought to be analyzed in a dispositional sense, I am not always sufficiently clear about whether I am analyzing the concept in dispositional terms. IR seems to leave open whether candidate substitutes include counterfactual ones, and even when I refer to possibilities, I’m not always explicit about what kind of possibility I have in mind. So, for instance, I write: “An object will satisfy IR and qualify as meaningfully irreplaceable if the fully specified criteria of replaceability are in fact such that no other object is or could be valuable in the same way—in other words, the object is the sole member of its substitution class,” (55) and, with respect to the umbrella, I write “If there were only one umbrella left in the world and no more could be produced, that umbrella would be an exemplar of irreplaceability: it would be the only thing valuable in precisely the same instrumental way that umbrellas are” (59).
    What I mean to be emphasizing with comments like these is that my analysis of irreplaceability is not counterfactual. This is part of the reason that it is so contingent: irreplaceability depends on whether other things with the same kind of value happen to exist, independent of whether there might be counterfactual replacements. So I agree, then, that there is a structural similarity between the case of the umbrella and the brooch: indeed, my aim is to treat all valuable objects with the same account. Whether or not either counts as irreplaceable, on my view, will depend on whether anything else is valuable in the same way. That’s why umbrellas are always replaceable in the world we live in, whereas mourning brooches usually are not. But that’s just how it happens to be.
    Now, to the case of persons. If there has been a unifying theme in this thread, it has certainly been that my account is missing something when it comes to persons. Rae helpfully brings in the Kantian distinction between dignity and price, and also notes that my interlocutors have been happy to accept that some extrinsically valuable things have a dignity, too. This suggests that they cannot merely help themselves to Kant’s distinction, since his is based on the distinction between conditioned and unconditioned value. So in contrast with previous concerns, Rae (if I am understanding her correctly) is pushing the Kantian line that only persons are properly regarded as irreplaceable.
    This is, in effect, another way of pushing for the conclusion of my paper: there is no important relationship between historical value and irreplaceability, but rather, irreplaceability is unique to persons. This approach has the great virtue, then, of ruling out the cases I am interested in arguing against while accommodating the case that gives my view some trouble. I myself am disinclined to take this route because I, like my interlocutors, want to preserve the intuition that non-persons can be irreplaceable too, though not merely (or only) in virtue of being historically valuable. Moreover, I still wonder, along the lines I have suggested above, whether it is not enough that persons are all, as a matter of fact, irreplaceable: whether we need the necessity of a sense of irreplaceability that would defy even counterfactual replacement, as in Kant’s notion of dignity. Is being in fact uniquely valuable an insufficient honor just because that uniqueness is contingent? I remain unsure, though I feel the force of the Kantian position.
    There is much for me still to think about in Rae’s comments, but I will close by registering agreement with the idea that treating people as having merely historical value would be a mistake. While it may be that a certain history is an essential part of a person’s value qua member of a relationship, it would be deeply objectionable to view a person’s value as consisting merely in that role. Rae notes the example of an anthropologist valuing persons merely for their history, and such an attitude has been forcefully critiqued, especially by those who have been the focus of anthropological concern (I recommend the chapter “Anthropologists And Other Friends” in Custer Died For Your Sins by Vine Deloria Jr.). My view may entail that certain relationships are replaceable, however unlikely that is to be the case, but it need not, I think, entail that the persons in those relationships are replaceable qua persons.

26. Thanks, Erich, that’s helpful on substitution classes. A bit of follow-up: you say that there might be evaluative reasons that constrain our reactions. But you also wisely point out that these reasons needn’t generate requirements. The attractive picture I think you have in mind is that we could both point out that child A is missing something in failing to treasure his deceased mother’s self-portrait, and yet the reasons generating this truth aren’t ones that compel us to criticize child B who likewise shrugs it off, given that *his* house is already stuffed with dozens of such artifacts.
    My only problem with this is that the reasons must generally be very, very weak, since as you say there is only so much treasuring, appreciating, valuing etc. to go around in one’s life. This does suggest that it will be very, very contingent whether in your particular case (given all the other random junk in your evaluative space) there are reasons that narrow the substitution class appreciably. Which would mean that the fact that something is irreplaceable to *you* doesn’t say that much either about the thing or you, since it’s all so contingent. It might be good to make the relativity here more explicit in these discussions.

27. Yes, Dan, that’s exactly the picture I have in mind, and it’s true that the contingency becomes quite radical in the case of mementos and keepsakes. I’m inclined to think that’s accurate to the phenomena, though. Take Cohen’s eraser, for example. I mean, what a random thing to latch onto, but he’s got a good story to tell about why he values it, and it may well be that little else meets that evaluative bill. But I would venture to say that Cohen’s story does tell us something important, both about the eraser and himself (but more about himself). I agree, though, that being more explicit about this aspect of the view would be helpful: thanks for helping me bring it out!