Wednesday, January 19, 2022

Learning from Science Fiction

guest post by Amy Kind

Thanks to Eric for inviting me for to take a stint as guest blogger here at The Splintered Mind. I’ve been having a lot of fun putting together this series of posts, all of which will focus in some way on philosophical issues raised by science fiction.

Like Eric, and indeed like many philosophers, I read a lot of science fiction. I won’t try to sort through the many possible explanations for why philosophers tend to be attracted to science fiction, but I’ll highlight one such explanation of which I’m especially fond. As Hugo Award winner Robert J. Sawyer has noted, science fiction would be better known as philosophical fiction, as “phi-fi not sci-fi” (see the back cover of Susan Schneider’s Science Fiction and Philosophy). Kate Wilhelm, another Hugo Award winner, makes a similar point in her introduction to an edited collection of Nebula Award winning stories from 1973:

The future. Space travel, or cosmology. Alternate universes. Time travel. Robots. Marvelous inventions. Immortality. Catastrophes. Aliens. Superman. Other dimensions. Inner space, or the psyche. These are the ideas that are essential to science fiction. The phenomena change, the basic ideas do not. These ideas are the same philosophical concepts that have intrigued [humankind] throughout history.

In fact, we might naturally take these claims by Sawyer and Wilhelm one step further. Not only does science fiction concern itself with the kinds of issues and problems that are of interest to philosophers, but it is also thought to provide its readers with important insight into these issues and problems. By engaging with science fiction, we can learn more about them.

As intuitive as this idea is, however, once we start to think about it more closely, we confront a puzzle. After all, science fiction is fiction, and the defining characteristic of fiction is that it’s made up. So how can we learn from it? If we really wanted to learn about space exploration or robots or time travel, wouldn’t we do better to consult textbooks or refereed journal articles focusing on astronomy or robotics or quantum physics?

I suspect that some think this puzzle can be easily resolved. The way that science fiction enriches our understanding is by providing us with thought experiments (TEs). Just as we can learn from TEs presented to us in philosophy – from Jackson’s case of Mary the color scientist to Thomson’s plugged-in violinist – we can learn from TEs presented to us in science fiction. The question of how we learn from philosophical TEs is the subject of considerable philosophical debate: Are they simply disguised arguments, or do they function in some other way? But the claim that we learn from them is widely accepted (though see the work of Kathleen Wilkes for one notable exception).

In his recent book Knowing and Imagining, however, Greg Currie has called into question this resolution of the puzzle. Though he is focused on fiction more generally rather than just on science fiction, his discussion is directly relevant here. In the course of an argument that we should be skeptical of the claim that imaginative engagement with fiction provides readers with any significant knowledge, Currie takes up the question whether one might be able to defend the claim that we can gain knowledge from works of fiction by treating them as TEs. Ultimately, his answer his no. In his view, there are good reasons to doubt that fictional narratives can provide the same kind of epistemic benefits provided by philosophical or scientific TEs. Here I’ll consider just one of the reasons he offers – what might be called the argument from simplicity. I focus on this one because I think consideration of science fiction in particular helps to show why it is mistaken.

As Currie notes, the epistemically successful TEs found in philosophy are notably simple and streamlined. We don’t need to know anything about what Mary the color scientist looks like, or anything about her desires and dreams, in order to evaluate what happens when she leaves her black and white room and sees a ripe tomato for the first time. But even the most pared down fictional narratives are considerably more complex, detailed, and stylized than philosophical TEs. These embellishments of detail and style are likely to count against from the epistemic power of the TE presented by the fiction. A reader won’t know whether they’re reacting to the extraneous details or to the essential content. Philosophical TEs would get worse, not better, if they were elaborated and told with lots of panache. And that’s just what fiction does.

In response to this argument, I want to make two points.

First, though epistemically successful TEs are generally simple, they do contain some of level of detail, and those details might well sway readers’ reactions – as Currie himself notes. Someone who has bad memories of their childhood violin lessons, or who associates violin music with a particularly toxic former relationship, might react differently to Thomson’s case from someone whose beloved partner excels at the instrument. Dennett’s TEs in “Quining Qualia” include lots of cutesy details, and so does Parfit’s description of his well-known teletransporter case. But despite this, we nonetheless think we can learn from these cases. Currie is undoubtedly right that we need to exercise care when engaging with TEs, and we need to be guard against being swayed by extraneous details. But in philosophical contexts, we generally seem able to do so – perhaps not perfectly, but well enough. So why wouldn’t that be the case in fiction as well?

Second, and here’s where consideration of SF becomes especially important, it’s not clear to me that simplicity is always the best policy. The seemingly extraneous details need be seen as so extraneous. Consider Andy Weir’s Project Hail Mary, and in particular, the character Rocky. Rocky is a sentient and intelligent alien hailing from the 40 Eridani star system. Members of the Eridian species do not have eyes and navigate the world primarily by using sound and vibration. In many ways, Weir is presenting us with an extended thought experiment about what kind of civilization such a species would develop. What would their interpersonal reactions be like? How would they make scientific progress? How would they achieve space flight? Trying to understand what it’s like to be an Eridian is a lot like trying to understand what it’s like to be bat – something Thomas Nagel has claimed cannot be done. But trying to understand what Eridian society might be like is not similarly out of reach, and Weir’s discussion helps enormously in achieving this understanding. Here’s a place where more complexity was helpful, not hurtful. To gain the understanding that I believe myself to have gained, I needed the fuller picture that the book provided. Without the details, I’m pretty sure my understanding would have been considerably impoverished.

While this is just one example, I think the point extends widely across science fiction and the TEs presented to us in this genre. Perhaps in other genres of fiction, Currie’s argument from simplicity might have more bite. But to my mind, when it comes to science fiction, the complexity of the thought experiments presented can help explain not only why philosophers would be so attracted to these kinds of works but also why we can gain such insight from them.

[image source]


Arnold said...

My question was-is before argument, science, fiction, science fiction, thought experiments...
...that eastern traditions have always the simplicity of our being here...

Where are we, as everything is in front of us, where am I...
..Is it, Socrates was teaching/seeking, one's own being here first...before philosophy...

chinaphil said...

Interesting. A couple of things occurred to me.
There's a sense in which we cannot "learn" from either thought experiments nor SF, in that they are both made up. The learning we do from this kind of literature must be more about reordering and better analyzing our own inchoate ideas.
One difference may be in the ideas that we bring to different kinds of writing. Usually a thought experiment explicitly asks you to imagine a situation as real. We are expected to apply our intuitions about the real world to its unreal situation. Science fiction, in contract, presents itself explicitly as a fiction (with some exceptions, e.g. the "found history" framing device in Handmaid's Tale and Dune); we generally approach fiction with a fairly fiction-specific set of reactions. To pick an obvious example, when I read about a space battle in science fiction, I am usually exhilarated. If I ever witnessed a real battle, I assume I'd be scared and sickened, and those are the intuitions I'd try to bring to a thought experiment battle. But I don't try to apply those "realistic" emotions to fiction.
(Speaking of space battles, anyone keep up with The Expanse? There was a great battle in the most recent episode.)

Ian said...

Hooray for direct discussion of philosophical methods and science fiction! It is hard to get tired of conversations on this topic, and philosophy needs more of it.

I think your second point is on the money-- science fiction extends an invitation to spend a while extending our empathic imagination in unfamiliar environments. Doing that work makes available insights that couldn't be had from a traditional 1-paragraph thought experiment. But I think that point suggests that philosophical TEs and philosophically interesting SF are after different goals and useful for different things. They are not just shorter and longer versions of a similar philosophical technique.

Very roughly, traditional TEs are most useful for challenging specific beliefs. (As, for example, Thomson's violinist challenges the belief that a right to life entails a right to life support.) I think Currie is basically right that extraneous details often risk weakening TEs, because the good ones should be carefully crafted to leave the reader with a belief formed in response to the TE that is unambiguously in tension with the specific belief the TE targets. (There is little question the belief that "I am within my rights to unplug from the violinist" is in tension with the belief that "a RTL entails a right to life support." If Thomson had padded that TE up to short story length, it would be much harder to form so specific a belief in response to it.)

Science fiction would be tedious if its goals were as constrained as that. Authors doing it well are imagining unfamiliar interactions between values and environments that can allow us a new perspective on our own values. TEs and SF can both put us into a state of disequilibrium, but TEs are more like surgical tools targeting a single specific belief, while SF is more like a travel experience that offers us a new perspective on our own values and understandings of the world.

Anyway! That's my almost certainly failed attempt to express in two paragraphs my current state of thinking about the philosophical value of SF. If you'll forgive a bit of shameless self-promotion, I have a paper in Metaphilosophy on this topic, sveltely titled "Stable Strategies for Personal Development: On the Prudential Value of Radical Enhancement and the Philosophical Value of Speculative Fiction." It's a two-part paper, and part 2 is a methods paper arguing for a use of science fiction distinct from philosophical TEs.

Amy Kind said...

Hi chinaphil -
Thanks for your comment! There is indeed something very puzzling about the idea that we can learn from something that's made up -- but I think we can and do. Consider an example that often comes up in philosophical discussion of imagination: If you are trying to make a decision about which of two sofas to buy, you might imagine how each one looks in your living room, and those imaginings help you to come to a decision. The imaginings are "made up" but you still can learn from them. Now of course you might be bad at imagining, so some imaginings (like some fictions, and some thought experiments) are not good ones to learn from. But, alas, some textbooks and lectures are bad, yet we can still learn from textbooks and lectures. I might try to return to these issues in a future post during this guest blogging stint; we'll see how it goes.
PS - LOVE the Expanse.

Amy Kind said...

Hi Ian -
Thanks so much for your comment. I will definitely check out the papers you mention. I'm glad to learn of them.
I like the "surgical tool" vs travel comparison. That seems like a useful way to think about things. And I agree that SF aren't just philosophical thought experiments. They aim to do more than what philosophers (or scientists, for that matter) aim to do with thought experiments. That said, I do think science fiction can challenge specific beliefs, and that it often does. Eg, Asimov's Bicentennial Man might be seen as targeting the specific belief that robots cannot be persons.
I also think that, though there may be some risks in adding more detail, those details can also work to help us engage more deeply and effectively with the TE. In this sense, it's a mistake to treat them as extraneous.

Luke Roelofs said...

I also like the scalpel/travel analogy, but maybe it's useful to think about them as extremes on a spectrum. A story like Bicentennial Man might have a specific target but unavoidably bring potential illumination on many other overlapping topics, sort of like a trip (pilgrimage) to see a specific location, where we can't help but also notice all the other places we're passing through. Making a thought-experiment more elaborated is like supplementing our surgical tools with a wider range of diagnostic and therapeutic tools to help ensure that we've identified the right problem and are optimising the outcome.

(...I was hoping to find a way to smoothly transition from one metaphor to the other but now I'm not really seeing it...)

Amy Kind said...

I think the spectrum point makes a lot of sense; thanks Luke!

Ian said...

Thanks for the reply, Amy!

I've been thinking about what could distinguish extraneous details that weaken a TE from those that don't. There's some sense in which details like "famous violinist" and "Society of Music Lovers" are extraneous details in Thomson; some philosophers wouldn't have included them. But it seems to me the violinist TE is better for having them.

(Another of my favorites, and one that seems to land especially well for students, is Norcross's "Fred's Basement" story from "Puppies, Pigs, and People," which is LOADED with extraneous details that contribute to its effectiveness.)

On the other hand, I've seen tons of bad extraneous details from students who are learning to write described cases. They tend to insert details that undermine their effectiveness. And I've definitely seen philosophers publish TEs with similar weaknesses.

One possibility is that is more a matter of good writing than philosophical methodology. When extraneous details efficiently contribute to a clearer picture in the mind of the reader or make the reading experience more engaging they are good; when they muddy the picture or are tedious they are bad.

Another possibility is that described cases used as counterexamples are usually better when they are minimal, while described cases used for an argument from analogy are more likely to benefit from details that bolster various aspects of the analogy. (This possibility might be in line with Luke's spectrum suggestion.)

I don't know! But thanks again for providing the impetus to think about it.


Philosopher Eric said...

I’d like to offer a rousing cheer for thought experiments! I’ve heard various distinguished intellectuals criticize them essentially as “fake experiments”. It’s a transparent move when some of their own beliefs may be so ridiculed. Furthermore I think we should remember their relevance to science itself. Here’s a post from Sabine Hossenfelder which classifies thought experiments as Einstein’s greatest legacy (rightly or not).

Furthermore what exactly do we mean by “thought experiment”? All I think we mean is that when someone proposes an idea about how things are, we can then explore that idea’s implications through various novel situations that an author might not have considered. I find it quite entertaining to ask the supporter of a dubious idea how comfortable they are with various implausible implications of what they believe.

Simple thought experiments do seem more effective than elaborate ones though I don’t consider this to make sci-fi a bad medium in itself. Regardless of how elaborate a given sci-fi scenario happens to be presented, an intelligent reader ought to be able to reduce the situation down to certain implications of what’s being proposed, or at least if a reasonable thought experiment does exist below.

I feel like I’ve become more comfortable publicly discussing the sometimes repugnant implications of my own ideas. If consciousness is fueled by means of a value dynamic as I believe, this suggests that as we build machines which can take care of us autonomously, we should also progressively substitute amazingly rewarding “fake” experiences for “real” ones (since engineered experiences might be many orders better than what’s standard). And my response to the human ultimately going full “blue pill”? Sometimes what’s true should naturally conflict with what we’d like to be true. Objectivity demands us to consider how things work from as large a perspective as we’re able to manage.

Amy Kind said...

Hi Ian, Those further thoughts seem useful. Glad to have prompted them! And of course, I like this possibility that you mention, that it is "more a matter of good writing than philosophical methodology. When extraneous details efficiently contribute to a clearer picture in the mind of the reader or make the reading experience more engaging they are good; when they muddy the picture or are tedious they are bad." If that's right, then that would help to explain further why we can learn from the thought experiments of (good) science fiction.

One last comment: In this case, it is a bit strange to call the details *extraneous* -- after all they're efficiently contributing to a clearer picture... So perhaps part of the point is that we need a better understanding of what counts as extraneous and what doesn't...

Amy Kind said...

And this last comment I just made to Ian also picks up on the helpful point from Philosopher Eric (thanks!) that even when thought experiments are elaborate an intelligent reader can still reduce them down to essentials... So now we might ask: well, then, wouldn't it better if the TE were just presented in that reduced form, with just the essentials? It's on this point I guess that I've been saying no. The very process of reduction might well be critical to the insights. And so again, perhaps it's a mistake to see those details as really extraneous after all.