Transcript

• "Grasping the Invisible": Analogies in Premodern Islamic Thinking and Beyond

  Hannah Erlwein: Our thinking really happens in analogies. So, I find it personally very interesting to think about in what different ways are we using concepts.

  And, of course, when you then look at a different tradition like the pre-modern Islamic context, you can ask the exact same questions.

  Stephanie Hood: Science Social, a podcast series about how science, history, and society connect with and add to the big questions that we all have today. This show is created by the Max Planck Institute for the History of Science. My name is Stephanie Hood and in each episode I’m joined by guests from our institute to talk about their research, their big questions, and some of the weird and wonderful experiences they’ve had along the way.

  ---

  Stephanie Hood: I’m really pleased today to welcome Hannah Erlwein, who’s a postdoctoral fellow in our Max Planck research group on Experience in the Premodern Sciences of Soul & Body ca. 800–1650. Welcome Hanna.

  Hannah Erlwein: Thanks so much for having me. I’m very happy to be here.

  Stephanie Hood: We’re here today to talk about analogies in Islamic thought. So, I wanted to start this podcast and this interview with you with an analogy or giving some examples of analogy. So, the dolphin looks like it’s smiling. I mean, is that an analogy or what is an analogy?

  Hannah Erlwein: Well, that’s a really good question actually. Because when you look into the literature, you will find that, apparently, it’s very difficult to define analogies. Maybe what we could say generally about how analogies work is that it transfers a judgment about one thing to another thing based on a similarity they share. And so, what we have in this particular analogy is that we’re basically drawing an analogy between humans and a dolphin. And the judgment that’s being transferred is smiling. And why are we transferring this judgment based on a similarity, which is in the case of humans that the mouth has a certain shape when it smiles? And then we’re looking at the dolphin and we’re recognizing a similar shape. And so we’re saying: “Oh, it must be smiling.” But again, as I said, the interesting thing about this sort of analogy is that it’s figurative in a way because, of course, we know that dolphins aren’t really smiling, but we’re kind of anthropomorphizing them. So, we’re establishing some similarity to how we talk about humans. And even though we know that dolphins aren’t really comparable to humans, we do draw an analogy.

  Stephanie Hood: That leads me a little bit to my next question actually… It’s just… I was wondering what got you interested in your research area on analogies in Islamic thought in the first place?

  Hannah Erlwein: I mean, my background is in Islamic studies, so I have generally an interest in Islamic theology and Islamic philosophy. So, one thing I noticed in reading those theological texts I engage with is that they are making use of experience quite a lot. So, they’re invoking experiences in order to gain knowledge about things that are beyond experience. But the interesting thing is they’re doing it by way of analogy. So, they’re drawing analogies between those experiential phenomena and these other phenomena that are not accessible by experience. So, I’m interested in asking the question: How in a given science, such as classical Islamic theology, you can actually gain knowledge about things. And so, yeah, in that sense, analogy is a really interesting topic.

  Stephanie Hood: I was actually back to ask what part of your research is the most exciting part… most exciting for you? Would that be it, do you say?

  Hannah Erlwein: I actually have to say what I found most interesting about all of my research is that it is an exercise in critical thinking. I find, you know, when you look at these theological or philosophical texts in a pre-modern Islamic context you will see that they were really trying hard to question assumptions. So, this is what I really find exciting about it. It makes you think about your own assumptions and ask questions about why do you believe what you believe, and do you have good reasons for it.

  Stephanie Hood: We’ve covered a little bit about your research experience. Now, I really wanted to talk about your research itself, and especially about how science, if we can call it that, looks in the medieval Islamic world. First of all, I mean, what kind of questions were scholars concerned with?

  Hannah Erlwein: Yeah, so I think first of all, you’re making really good point when you’re saying if we can call it “science” because, of course, there’s quite a big debate going on, of course, among historians of science. Whether we can actually use this term to refer to other traditions of how should we call them “knowledge making” or something like this. I’m personally much more interested in looking at how historical actors in different time periods but also traditions understood what they were doing then. I think maybe the way we should understand science is more in a general way of a systematized way of knowledge making, which is different from, you know, everyday knowledge and things like that. So, in that sense, I would actually speak of science in the medieval Islamic context, a pre-modern Islamic context. But yeah, you get sciences such as philosophy, theology, law, mathematics, astronomy was a big thing and, yeah, all kinds of other sciences.

  Stephanie Hood: On going from science into religion, how would you say the relation between religion and science looked? Do you have examples?

  Hannah Erlwein: I think with some of these branches of science, it’s very evident that they have an actual relationship to religion, such as theology. But even in that case, what’s actually interesting is some of the topics which theologians discussed, you could say, they’re not in themselves related to God–only in consequence they’re related to God. So, for example, they were inquiring about the makeup of the world. So, they were asking what stuff is the world made of and then gave the answer it’s made of atoms that form bodies. Or, for example, they were discussing what knowledge means, how do we acquire knowledge. So that’s not in itself a religious topic as such. But in being part of the science of theology it became a religious topic because they were using all these discussions to eventually move on to asking questions about God and the world’s relationship to God and so on. But then we also have some other sciences, such as, say, physics or astronomy, where you can definitely see that they could kind of do without being immediately connected to religion. However, in that case, you could also argue that they use this kind of knowledge for religious problems. For example, when they were trying to calculate the direction of prayer. So, you know, in Islam, all Muslims have to pray towards Mecca. And, of course, you need to then calculate which direction you have to pray. And so they used astronomy and mathematics for that. And so, in that sense, these sciences, which in themselves are not really religious, they became used for religious purposes. So, it’s a really complicated question about the relationship between religion and science, but I think it’s a very interesting one.

  Stephanie Hood: There’s something that you write about in your research and you also wrote a feature story for us recently, the way you touched on this topic of kalām. Actually, what is kalām? What’s it concerned with?

  Hannah Erlwein: In the literature, you will always find kalām being translated as speculative theology. The word kalām in Arabic actually means a conversation or speech or debate. And I think this is actually indicative of its origins or like how the practitioners of kalām understood their own science, what it was about. Because a big part of kalām is to present rational proofs for religious dogma. So, the way the mutakallimun, and that’s the word for the practitioners of kalām, the way they understood their own science was that they said, well, it’s not enough to believe something simply by following religious authorities. So, it’s not enough to say, I’m a Muslim and I believe in God and Muhammad as his prophet and so on, simply because my parents or some religious teachers told me to do so. Because they argued that if you follow one person without proof, then why not follow another person who also doesn’t have a proof? And so they said, in order to follow, for example Muhammad, who claimed to be a prophet from God and so on, you have to have a good reason. And what’s interesting is they did this both in defense of Islam against other faiths. So again, here comes in this notion of kalām, meaning a debate. So, for example debates with Christians or Jews or whoever else they encountered. But it also means defending doctrines internally. Because, of course, Muslim theologians did not all agree on the same dogmas. And so they were debating among each other as well.

  Stephanie Hood: What kind of… so you mentioned something about the debate… what would that have looked like?

  Hannah Erlwein: Yeah, so we know that historically, these theologians were actually debating with each other, so they had actual debates like we’re having a conversation now, so to say. But what’s maybe more interesting for us is actually looking at the works that have come down to us. And so, what you will find in there is that since the early beginnings of the science of kalām, these works are structured in a specific way. So, for example, you get this thing where a theologian assumes a hypothetical interlocutor or some kind of opponent and basically uses this opponent in his work as you could even say a rhetorical device in order to, for example, introduce a given problem and then to answer to it, or in order to maybe pose an objection–a possible objection–in order to refute it. And what we can also see from the works is that these Muslims were not just concerned with “Oh,” like, “now I'm going to explain to you” or “I'm going to give you the proof for Muhammad’s prophethood” or something like this. But that they were doing this actually with actual opponents in mind. One example is there’s an early theologian called Maturidi, and he lived in Samarkand, so Central Asia, today’s Uzbekistan. And in Uzbekistan, he actually did, in reality, encounter Zoroastrians. So, these are religions, you know, those dualistic religions where they don’t believe that there’s this one God who created everything, but they believe that there’s two principles, good and evil, that created the world and these kinds of things. And, of course, for a Muslim like Maturidi, he actually did engage with these people in debates on these theological issues. And he does so in his work as well. So, when you look at who are the opponents in his work, either they’re more Muʿtazilī theologians, so theologians from a different group, or they’re adherents of these kind of dualistic religions.

  Stephanie Hood: So, what kind of experiences were considered to work as analogies or what kind of analogies were thought to indicate something about phenomena that were beyond sense experience. That’s also something that you’ve written about.

  Hannah Erlwein: Exactly. So, if we’re now focusing here on the branch of science that is theology, they had different names to refer to this sort of analogy. But, for example, what they called it was al-istishhād bi’l-shāhid ʿalā al-ghāʾib, which means using the shāhid as evidence of the ghāʾib. And these are two important terms. The shāhid refers precisely to all the phenomena that are known by experience. And the ghāʾib refers to all the phenomena that are beyond the reach of the senses. But literally what these terms mean is shāhid comes from a root that means testifying to something or being present, whereas ghāʾib comes from a root that means being absent. And so this was kind of the notion of analogy they used. So, what kind of particular experiences did they use? So I would say with the theologians, there were certain experiences they all liked to use and which apparently they agreed on that they’re good cases of an analogy for those unexperienceable phenomena. So, for example, they made use of analogies–things like “buildings” and “builders” or “writings” and “writers”. So, what they would argue, to give an example, is that they said: “We know that buildings only come about because of a builder” or “Every writing that comes about has an author.” And then they would use this as an analogy for the origin of the world. And so, yeah, they drew an analogy between how buildings have a builder, an analogy that said the world, which also has come about, needs to have a cause which is God. Which, again, belongs into this realm of the unexperienceable phenomena.

  Stephanie Hood: Are there also, would you say, European historical equivalents for such uses of analogy in scholarly work?

  Hannah Erlwein: I’m actually not sure. And this would be a really interesting research question, so thanks for pointing it out to me. So, I’m not sure about to what extent in the Western context of science they adopted analogies they would have found in the Islamic context. But what’s definitely the case is that in the context of Western science, we also find a lot of analogies. And, for example, one of them would be this Cartesian worldview of basically in this specific context in the 17th century. They would have an affinity for drawing analogies between machines and humans. So, another good example would be Newton and his famous analogy between the surface of the moon and valleys on Earth. So apparently, Newton was observing through the telescope that, when he was looking at the moon, that he would find light and dark spots there. And he sought an explanation for that. Because back then there was this belief that the heavenly bodies are perfect spheres. And so, a perfect sphere, of course, cannot have valleys and mountains and so on. But he tried to explain how it is then possible that he sees those dark and light spots on the surface of the moon. And in trying to explain this, he actually drew an analogy to the surface of the Earth, where we also observe this phenomenon of light and dark spots precisely because the surface of the Earth has valleys and it has mountains. Ok, so this example of Newton is also interesting because what precisely can an analogy actually do? So, for example, does an analogy in science work as actual proof or is it more something useful for actually coming up maybe with hypotheses? And so, when it comes to Newton’s analogy, the question then is, well, is this analogy that he’s drawing to the Earth, is it enough as a proof of what the surface of the moon actually looks like?

  Stephanie Hood: That’s a really good point about the danger of false analogies, I think, which is something I wanted to ask you. So, if analogies were used to prove unobservable phenomena, and there was, of course, a risk that some of these were false, how did the Muslim scholars deal with that?

  Hannah Erlwein: This is something my historical actors were actually very much concerned with. And so maybe I can give one example here to illustrate this kind of debate. So, there was one early theologian by the name of Ashʿarī, and he made use of precisely these kind of analogies I mentioned earlier. An analogy to this experience that buildings have builders. He concluded that the world must be created by God. And so there was one school they were called the Muʿtazilītes. So, they reminded the followers of Ashʿarī that Ashʿarī subscribed to a theory of causality, which denied that humans actually create their own actions. So, what does that mean? So, imagine I’m taking this glass of water that’s standing in front of me, I’m lifting it. So, I think this is me who is doing it. But according to Ashʿarī, this is an illusion because, for him, only God is the only real agent and cause in the universe. So really, it is God who creates in me this action of lifting the glass of water. What is the consequence of this for this use of analogy between buildings having builders and then the world therefore needs to have a cause? So, these Muʿtazilītes said: Well, Ashʿarī, you don’t actually think that these buildings come about because of the builders? Because you deny that the builders are the actual agents bringing about these buildings. In reality, according to your theory of causality, it is God who is doing this. So, if humans don’t actually bring about buildings, you cannot use this as an analogy to ask them this question about does the world have a cause? Is God really the creator of the world? Because how can you use the observable realm as an analogy to the unobservable realm in order to establish the principle of causation if you deny that what you’re observing is an actual case of causality? For them, what they were criticizing is that you’re using these analogies, but they’re false analogies because according to you, they don’t actually work. You can’t use them. And so, these Muʿtazilītes, of course, could put forward this critique because they subscribe to a different theory of causality, which was the precise opposite that they said: So, it isn’t God who creates this action. It is really me who does this. And so, for them, subscribing to the right theory of causality, namely that humans actually bring about the actions, was a prerequisite or condition for saying that drawing an analogy between humans as actors and God as actor is actually valid.

  Stephanie Hood: Wooah… I actually really understood it, but at the same time, my brain is doing flip-flops.

  Hannah Erlwein: See, this is actually the exciting thing. This is also what I meant earlier when I said this, like, rethinking why we actually hold assumptions. Because when you suddenly start to think about analogies, it’s like, we all use analogies. But then it’s really this question, like, well, is it a valid analogy? Like, why are you supposed or allowed to draw an analogy between two things? It’s a bit like in the history of science. There’s this big discussion about the theory-ladenness of observation, right? Where they basically say, when you’re observing something, can you ever see the thing itself? Or is it not rather the case that our theory works a bit like glasses? And so, depending on what glasses you’re wearing, you’re seeing different things, right? When I give this example of… if you take Kepler and what’s his name, Tycho Brahe? So, one believed in geocentrism, the other one in heliocentrism, right? And when you put them on a hill and have them observe the sunrise, would they actually see the same thing? Because of course, depending on what theory of like the solar system you have, you will of course describe what you’re seeing in a completely different way.

  Stephanie Hood: That actually brings me perfectly to my next question I wanted to ask, which is about what function analogies actually have. So, would you say that they were proper arguments or were they just intended for persuasion?

  Hannah Erlwein: And this precise same question was asked by my historical actors. So, what can analogies actually do? Are they simply good for maybe like putting forward hypotheses in science? Or can they actually prove something? So, in the context of kalām, when we look at the early generations of theologians, the ones I just mentioned–like those Ashʿarī, those Muʿtazilītes and so on–they thought that these analogies they’re using actually are a proof that they’re an actual way to knowledge. Because remember, I said earlier that what the science of kalām was built on was to say: You need to find a proof for religious dogma; you cannot just believe it just so, so to say. But the interesting thing then is, and this is also part of my whole research project, this conception of analogy did change later, which is exciting because there was one very famous philosopher, I mentioned him earlier, Ibn Sīnā, also known as Avicenna. Ibn Sīnā really stood in this Greek tradition of logic, some Aristotelian tradition. When you read through what he has to say, you can clearly see that he has the theologians in mind and their kind of analogies, the ones they’re using. And so, he is very critical of analogy because for him, following the Aristotelian tradition, the only way of gaining scientific knowledge, you have to have a demonstration. So, it’s a kind of different form of proof, so to say. And analogy can never live up to that. So, the conclusion Avicenna–or Ibn Sīnā–reached was that analogy is good in rhetoric. He says it’s often used by theologians and politicians because it’s a good way of convincing the audience. But he doesn’t think that analogy belongs into the realm of science. So, in this later philosophical tradition, you will find scholars, theologians basically, who also follow Ibn Sīnā’s critique and say, no, analogy doesn’t really work. We need other forms of argument. So, to illustrate the extent of this, I mentioned earlier Ashʿarī, and this Ashʿarī was actually the founder of a very important theological school. And so, it is quite something for later Ashʿarī theologians to break with whatever the founder of his school did. So to say, this founder of my school used analogies and he thought the work is proof, but I don’t actually agree with that. That’s quite something actually in intellectual history, you can say.

  Stephanie Hood: Would you say then that analogies are used in science today?

  Hannah Erlwein: Scientists nowadays also still do use analogies. But as I said, the question then is, what is the precise purpose? And I think probably what we find in modern science is more this point Ibn Sīnā was also trying to make: Analogy can help us come up with new ideas, new hypotheses in science, but it can never be a proof. You need to have some other method to then prove the hypotheses you came up with through analogy.

  Stephanie Hood: Going on to the last point that I wanted to chat with you about… about culture and communication and the wider context of your research… So, would you say that analogies can also be culturally specific, maybe influence our cultures?

  Hannah Erlwein: Oh yeah, definitely. I mean, this goes also back to what we talked about earlier with, for example, these Cartesian analogies. I mean, there was a reason that these analogies developed against a certain philosophical or intellectual background, which has to be there in order to draw these kind of analogies. But I think what you could also say again is the history of a country, which is not shared by other countries, can be used in order to draw analogies. So I mean, in the sense of that one draws analogies between historical events in the past and our present situation or something like that. And so, of course, if the history of a certain country differs from another one… and in one context, you can draw these analogies and other you can't. And so, in that sense, the specific analogies humans have made recourse to in history were very much, I think, reflective of their cultural background, of their religious background, their intellectual background and so on–yes.

  Stephanie Hood: Analogies, especially if we use the term in a broad sense, seem quite essential in the way that we as humans think and communicate. I wonder how much they might relate to science and technology that we develop today, even something like artificial intelligence. Would you say that there’s a connection there?

  Hannah Erlwein: Well, first of all, I agree with your judgment that analogy is really essential to the way we think. And this whole research area of machine learning obviously comes from the knowledge we have about how humans learn. And, of course, people have–or researchers have–for a long time recognized that we very much learn based on analogy. So, this whole analogy business does actually play a very important role in the development of machine learning and artificial intelligence. I actually just watched a really interesting video where an Oxford scholar, his name is Nick Bostrom, he was talking about the development of artificial intelligence and the question whether there could be machines one day that are more clever than humans and so on. And so, he made a very interesting point, which, of course, I took notice of since in my research I deal with analogies. So, he talked about how in machine learning, for example, you know, machines, they are taught to play a certain game or something like this, like a computer game. And they can get better at playing a specific game the more they play it. And then you feed it a second game and the machine simply doesn’t know what to do. What you have to do then, as a scientist, is feed this machine again new information. So based on this new information, it then can kind of get used to the second game. You would have to do this with all the other games as well. Because it cannot say: “Oh, I’m recognizing something in this game that I’ve already done in this other game.” All it does is: new game. So, this transfer of knowledge is precisely based on the ability to recognize analogies between one thing and another. I mean, think about it, for us humans, we learn how to do one thing, for example, and then we’re put into a new situation. But somehow we recognize that there are similarities between this new situation and the old one. And so, we can make recourse to what we have learned about how to deal with the first situation and therefore also then deal with the second situation. And apparently, when it comes to machine learning, this used to be for a long time a real problem. But really, what scientists are trying to do is, they’re trying to make use of what a machine has already learned in order for it to be quicker in learning a new situation. One example I read there was as well, say, a machine knows how to automatically drive a car and then you want the same technology for driving a truck. Of course, there’s differences between the two of them. A truck is not the same as a car in terms of navigation or something like this. But still, they’re similar enough so that you could hope that if the machine knows how to drive a car automatically, you can benefit from this knowledge and basically help it understand quicker how to drive the truck. And as I said, this is precisely the case of transfer of learning by analogy.

  Stephanie Hood: This is not quite the same thing but I just find it kind of interesting that there was this whole thing with Google Translate recently that it basically, how it’s reinforcing existing gender stereotypes because all the information is basically drawing out of… that, you know… from where it’s getting all of its information or of all of its language and translations. So like doctors are men and like nurses are women.

  Hannah Erlwein: Yeah, and I think this is another example with these machine translations because again it works on analogy. So, what I understood from this technology is that, I mean, as we know, words in different contexts mean different things. So that the correct translation depends on the context in which a word appears, right? So, the way these machines learn how to translate is by people feeding them loads of examples of a word in a certain context. But, I mean, this is just statistics, right? And then you will have maybe lots of cases, especially when it comes to translations of, let’s say, poetry, which is like a business of its own, which is much more difficult than translating an instruction of how to start a washing machine or something like this. Then the machine really has difficulties because it can only operate on those statistics but it cannot really, you know, get a sense of an intention and a text or like: On what level are we talking? On the literal sense? Or the metaphorical sense, for example.

  Stephanie Hood: Wow, this is fascinating. Now I really want to listen to your podcast…

  Hannah Erlwein: … Soon! (laughter)

  Stephanie Hood: … No pressure. Is there anything else that you think that we could learn from the use of analogies in kalām today?

  Hannah Erlwein: What I would just say is where do we use analogies? We use them literally everywhere. So, in everyday speech we use them. Politicians use analogies. Analogies are used in culture. They’re used in poetry, in everything. So, what I find really exciting about this whole analogies “business” and problem and looking at how in different time periods and cultural contexts analogies were used also in relation to science and what status it had was that I think it tells us something about us as human beings and about the way we think and how we basically strive to really gain knowledge. And I’ve talked already quite a bit about how important I think it is to think critically and question assumptions. And also nowadays where we’re confronted–especially through social media and so on with many different opinions and so on–it’s even more important maybe to be conscious about why we think the way we think and whether we have good reasons for these kind of things. So yeah, I guess that’s why it's important to look at analogies in pre-modern Islamic theology.

  Stephanie Hood: Awesome, awesome, awesome. Okay, just rip that mic off and drop it. That was a really cool end. I think maybe I can just finish with a plug for the podcast…

  Hannah Erlwein: …Yes.

  Stephanie Hood: I think there’ll be plenty of people listening who are now really becoming interested in the topics that we’ve discussed and you actually have a podcast coming out.

  Hannah Erlwein: Yeah, thanks so much for mentioning that. So actually, I’m going to soon hopefully start a podcast of my own, which is called Kalamopod. So, Kalamopod, of course, taken from this name kalām. I’m really hoping to introduce these themes in classical Islamic theology, which as we saw are very exciting themes to a general audience. And I would, of course, be very happy if many of the listeners of this podcast were also going to listen to my own podcast, which hopefully will actually be launched by the end of April. So pretty soon. And which, of course, is available on the usual platform Spotify, iTunes, AnchorFM. And I’m also going to upload videos on YouTube.

  Stephanie Hood: Great! I’m looking forward to listening.

  Hannah Erlwein: Thank you.

  ---

  Stephanie Hood: This is it for today. If you like what you’ve just heard, we love your support. Click the subscribe button, recommend this to your friends and colleagues, or give us a thumbs up in your favorite podcast app. You can find us on iTunes, Spotify, and anywhere else you can listen to podcasts.

  Science Social is produced by the Max Planck Institute for the History of Science. Music by Poddington Bear, then I’m the host Stephanie Hood.

  Make sure to follow us on Twitter at @MPIWG. And most of all, thanks for listening.