My first substantive post for this blog was a review of one of my co-contributor’s books—Derek Turner's Paleontology: A Philosophical Introduction (2011). Today I’m going to discuss the book of another co-contributor—Adrian Currie’s Rock, Bone, and Ruin: An Optimist’s Guide to the Historical Sciences (2018). Leonard, whenever you have a book, you’re totally owed a review of it from me! Just say the word—fair is fair. :-)
How cool is it that we get to discuss each others’ work in this forum? I haven’t shared this review with Adrian before posting, so I’m really excited to hear what he thinks about what I think about his book. Of course I think it’s a super fun read. Just chock full of great ideas and detailed case studies and provocative arguments. Adrian, unsurprisingly, doesn’t disappoint.
Because the book has so much going on in it, and because I don’t feel beholden to certain norms of comprehensiveness that might come with a more official review for a schmancy official publication, I’ve going to focus this review on two particular aspects of the book that I think are especially interesting, rather than attempting a more thorough assessment. I think this venue provides a unique opportunity to be more informal, frank, and specific than usual when reviewing and I want to take advantage of that. Of course Adrian has no obligation to respond to anything I write here! But I’m hoping to start at least a little bit of a conversation.
I’ll begin with something cute, since Adrian’s book is itself so cheerful and endearing. He describes writing the book while in Calgary and Canberra—and I read and reviewed his book while staying with a friend in LA. So here are some samples of local rock, bone, and ruin from where I’m currently at:
On the left, Heizer’s Levitating Mass at LACMA. In the middle, a small piece of the massive wall of dire wolf skulls found at the La Brea Tar Pits & Museum site. On the right, ongoing excavation efforts at the totally ruinous tar pits.
Ok, I have two things I want to discuss. One is about a confusion I have and the other is more contemplative. I’ll do the confusing one first, since it is potentially more critical. Then I can finish with something a bit more suggestive and fun.
The overarching argument of the book is that we should be optimistic, rather than pessimistic, about the historical sciences. Adrian lays out this aim right at the start. He introduces three “grounds for pessimism” and argues that all three “propositions underlying pessimism are false” (page 8). Here are the three grounds / propositions (again, this happens on page 8 of the book):
- Our available evidence about the past is limited to traces.
- Much information from the past has degraded or disappeared.
- Historical scientists cannot manufacture evidence.
Right away, when first reading this, I thought something like “ooh, 1 and 3 are prime targets! I can see why Adrian would want to take those on. But what about the middle proposition—claim number 2? How can that possibly be refuted?” It just seemed so obviously true to me.
As the introductory chapter went on, I looked for hints as to how Adrian might tackle claim number 2. I saw plenty about what was wrong with claims 1 and 3, but the critiques of claim 2 were more oblique. Adrian seemed to be targeting what someone like Lewontin (1998) has historically inferred from claim 2, rather than claim 2 itself.
The idea of three reasons to be pessimistic reappears in chapter 5, where Adrian identifies these as “three targets for the rest of the book” (page 134). But the list of reasons supplied here (on page 135) is slightly different than the one supplied earlier, in the introductory chapter:
- Our available evidence about the past is limited to traces.
- We are unlikely to uncover further traces.
- Historical scientists cannot manufacture evidence.
A related list of “three claims that, if true, would drive pessimism” (page 277) reappears near the end of the book (also on page 277):
- Our available evidence about the past is limited to traces.
- Historical scientists cannot manufacture evidence.
- We are unlikely to uncover further traces.
The order of claims here is different but I don’t think that matters at all—otherwise this third list is entirely consistent with the second list (the one provided on page 135). All of the claims on both of these later lists seem well handled by Adrian to me; I think that he does a compelling job in the book of undermining them and hence these three grounds for pessimism.
But what I’m really interested in is what happened to claim number 2 from the first list (the one provided on page 8). Putting all three lists together we’ve actually got four grounds for pessimism, since the original claim 2 (“much information from the past has degraded or disappeared”) is itself missing from the later lists, to which a new claim has been added (“we are unlikely to uncover further traces”). I’m not sure I’ve seen a conclusive refutation, anywhere in the book, of that original middle proposition. And this time it does matter to me—because I’m worried that this particular claim is one that provides enduring grounds for at the very least agnosticism about our chances of attaining knowledge about certain aspects of the past.
Adrian’s overarching argument for optimism hinges on (a) undermining the grounds for pessimism and (b) doing so for even the so-called “unlucky” epistemic cases. Part (b) is necessary because just succeeding at part (a) by itself doesn’t actually get us all the way to optimism. Pessimism in this context is thinking that we often don’t have much chance of uncovering past events. Undermining the grounds for that sort of pessimistic thinking doesn’t actually push us all the way towards optimism, since there is an available agnostic position between the two poles of pessimism and optimism.
I think of the pessimistic position as a position of confidence about our general inability to recover past events, whereas the optimistic position is a position of confidence about our general ability to recover past events. But we might think that really, we shouldn’t have any sort of confidence about our general (in)ability to know about the past. This is the agnostic position—it’s one that says, actually, we just can’t tell in what direction we’re trending here. Certainly, we add to our knowledge of the past all of the time. But as Derek points out (in Turner [2016]) and Adrian acknowledges (on page 283), adding to our knowledge of the past often also increases the number of things that we then know we don’t know.
So, how do we identify any sort of trend in this clouded context—are we generally able to recover past events, or are we generally unable to recover past events? This uncertainty is what supports the agnostic position: there is at least enough opacity when it comes to the past to obscure the nature of our access to it, yet we are successful enough in recovering it, sometimes at least, for us to also be unsure about whether and how often we’re really missing anything significant. (For those interested in an in-depth treatment of this sort of issue, I recommend Stanford [2006].)
Adrian tries to push us past the agnostic position and towards optimism by undermining pessimism even in the unlucky circumstances—those which supposedly provide the best-case scenario for pessimism. And I believe that his “ripple model of evidence” is supposed to help explain why, even in the unlucky cases, we have no grounds for pessimism. This is the focus of chapter 5, by the way, and it’s how part (b) from above is supposed to happen, I think. It’s just that the kind of cases Adrian uses to explicate his ripple model of evidence don’t seem to me to be genuine instances of what we might want to call the unluckiest of circumstances.
Consider the metaphor with which Adrian has named his model: the pattern of ripples that a thrown pebble generates in an undisturbed pond. He writes: “imagine I throw a pebble into an otherwise undisturbed pond and take snapshots of the resulting disturbance at set time intervals. Earlier times will have a smaller area of effect than later, and the disturbance may become more pronounced as time passes; however, in later snapshots the clear patterns generated by the pebble will distort and fade” (pages 111–112). Normal pebble-into-lake throwing activities aren’t like giant meteor impact craters—they don’t leave much of a record of their ripples. It’s the snapshots that we can imagine lasting here, and degrading over time, but still giving us something of a picture of what originally happened. When we imagine the ripples themselves lingering, to be variously degraded and preserved over a long time scale, I think we’ve started imagining something substantially different than what happens when a pebble is thrown into a pond—especially if no one snaps a picture.
Think about all the rocks at the bottom of any given lake. They all got there somehow, though at different times and in different ways. Which ones splashed through the surface of the lake, and what did each of the particular pattern of ripples that those rocks generated look like? Without anyone around taking snapshots of such an ephemeral process, I just don’t see us successfully recreating that sort of history with any kind of regularity, completeness, or confidence. I’m sure that the angles at which various rocks lie in the lakebed and the nature of the divots in the sediment nestled around them could tell me an unexpected amount—but what about the rocks that lie buried farther underneath the surface, with their angles and landing positions disturbed? What about the entry patterns of rocks lying in lake beds that have undergone serious disturbance, either due to plant growth or geologic activity or whatever else? Isn’t there a whole bunch of historical information that we will just never get?
I’ve arrived back at one of Adrian’s initial grounds for pessimism—claim number two from the first list, the idea that “much information from the past has degraded or disappeared” (page 8). I’m worried that pond-surface-ripple-history is precisely the sort of information about the past that we could not generally recover without someone around snapping photographs. By talking about just such a process (the ripples), but adding an extra evidence-generating component (the snapshots), Adrian has appeared to transform the worst-case scenario into a surprisingly accessible one. But the actual worst-case scenario is one more like the pebble and the pond without the picture-taker and the pictures. Or perhaps most like not even a rock at all but something more like an unseen bubble landing on the surface of a lake, and dissipating without any sort of ever recoverable trace.
Maybe those sorts of event are not supposed to be significant enough to matter, but I’m worried that we don’t know enough about them to know how much they matter. There have been a lot of such processes, and at least some of them might have mattered somehow. It’s hard to say anything especially confident or well-informed here—hence the agnosticism.
Anyway, this first issue was about a bit of a confusion I have, regarding what happened to that middle proposition from the first list of grounds for pessimism, and how that claim potentially underwrites what I’m calling the unluckiest of circumstances. I have lingering concerns about such circumstances and what they mean for overall claims about our knowledge of the past. Although Adrian has done a lot to push back against most of the grounds for pessimism that he introduces, there remains enough fertile opacity about the past left for me to stay agnostic, rather than to go as optimist as Adrian otherwise encourages.
On to the second thing! I really, really loved Adrian’s discussion of analogy and its virtues—a discussion that really gets going in chapters 7 and 8. Traditionally philosophers of science have, let’s say, privileged trace evidence over analogous evidence. (See Turner [2005, 2007] and Griffiths [1994, 1996, 2006, 2007] for exemplary articulation and defense of this sort of view.) Adrian bucks the trend here and offers a spirited defense of the evidential value of analogues.
I don’t plan to critically dispute Adrian’s account. Instead, I want to discuss the distinction between trace and analogous evidence, simply because Arian’s discussion has prompted me to think directly about that distinction in what is a new way for me. What precisely is going on with all this supposedly analogous evidence?
Adrian characterizes trace evidence as evidence of events unified by a joint history, connected via causal pathways. He says that trace evidence “is evidence by virtue of being relevantly downstream of the target—that is, being at least minimally dependent on one another in a fashion picked out by justified midrange theory” (page 198). In contrast, analogous evidence is evidence of events unified as types, connected as instances of the same kind of process. “Here, object and target are unified by virtue of being tokens of the same system type, instantiations of the same causal system. These systems are represented by models” (again, page 198).
There are some bits of evidence that are obviously traces in the sense of trace evidence described above. Adrian provides lots of examples in his book of cases involving trace evidence; probably my favorite is the startling use of LiDAR mapping to identify obscure physical traces of Mayan geography in aerially surveyed regions (see chapter 4). Other bits of evidence are obviously analogical, like those obtained via comparison of organisms occupying a “common niche” or sharing a “killing style” (both cases discussed in chapter 8). What about bits of evidence that share a causal origin, not due to an historical pathway connecting event A to trace B, but instead because they are shaped by something else that they have in common? Not something similar—that’s obviously analogy. Something the same. Relation or constraint C causes or influences both phenomenon D and phenomenon E.
What does it mean for two phenomena to be caused by the same causal system—not the same system type or same kind of process, but the very same cause, like a common constraint? To be subject to productive constraints that the two phenomena have in common? When Adrian talks (during his fantastic discussion of the Thylacosmilus atrox case) about the relationship between gape angle, bite force, and required strength, that relationship just seems like one dictated by physics to me. The relation that holds there is not something that applies analogously across different evolutionary trajectories—that just is a shared causal constraint that those organisms have in common.
I think that there might be an intermediate kind of case hidden between the obviously trace evidence cases and the obviously analogous cases. Joint history or shared causal pathway is one thing; same kind of process or same system type is another; but what about the very same causal constraint? Can’t two things have a common cause without those two things being connected by the same historical instantiation of that cause? Aren’t there at least a few causal constraints that hold everywhere, ones that can cause things to happen in ways that are not just similar but the same?
Basically, I’m suggesting that the crucial notion in play here—that of same / shared / similar—might have two importantly distinct senses: (i) a mere similarity of appearance or likeness, and (ii) a genuine similarity born of commonality. And that commonality also has two senses: (ii*) a commonality born of common trajectory or shared causal history, and (ii†) a commonality produced by common circumstances or shared causal environment.
Discussion of trace evidence generally focuses on sense (ii*). Discussion of analogous evidence generally focuses on sense (i). But some of what has typically been classified as analogous evidence seems to me like it might be of a different kind—the kind pertaining to sense (ii†). Am I making any kind of sense here? Not sure that I am. But if so, I think that the value of evidence corresponding to sense (ii†) might be greater than that traditionally afforded analogous evidence. So I’m not sure we should be lumping sense (ii†) evidence together with sense (i) evidence, into the somewhat besmirched general category of analogous evidence.
Well, I said that this part was going to be suggestive and fun—and I feel confident that I got the suggestive part down. Thanks to Adrian for exploring analogous evidence in a way that got me speculating so wildly! Alright, that’s it from me. You should read this book! Thanks for reading my post about it.
References
Currie, A. (2018), Rock, Bone, and Ruin: An Optimist’s Guide to the Historical Sciences (Cambridge, MA: MIT Press).
Griffiths, P. E. (1994), “Cladistic classification and functional explanation,” in Philosophy of Science 61: 206–227.
Griffiths, P. E. (1996), “The historical turn in the study of adaptation,” in British Journal for the Philosophy of Science 47: 511–532.
Griffiths, P. E. (2006), “Function, homology, and character individuation,” in Philosophy of Science 73: 1–25.
Griffiths, P. E. (2007), “Evo-devo meets the mind: Towards a developmental evolutionary psychology,” in R. B. Sansom & N. Roberts (eds.), Integrating Evolution and Development: From Theory to Practice (Cambridge, MA: MIT Press), 195–226.
Lewontin, R. C. (1998), “The evolution of cognition: Questions we will never answer,” in D. Scarborough & S. Sternberg (eds.), An Invitation to Cognitive Science, Vol. 4: Methods, Models, and Conceptual Issues (Cambridge, MA: MIT Press), 106–132.
Stanford, P. K. (2006), Exceeding Our Grasp: Science, History, and the Problem of Unconceived Alternatives (Oxford: Oxford University Press).
Turner, D. (2005), “Local underdetermination in historical science,” in Philosophy of Science 72(1): 209–230.
Turner, D. (2007), Making Prehistory: Historical Science and the Scientific Realism Debate (Cambridge: Cambridge University Press).
Turner, D. (2011), Paleontology: A Philosophical Introduction (Cambridge: Cambridge University Press).
Turner, D. (2016), “Another look at the color of dinosaurs,” in Journal of Studies in History & Philosophy of Science Part A 55: 60–68.