Michael Strevens and J.D. Trout on explanation and understanding in science (and beyond).
Do you prefer Youtube? Go here.
In science and in ordinary life, we want to explain and understand features of the world around us. How can we tell a good explanation from a bad one? What’s the connection between explanation and understanding? Strevens and Trout begin with an overview of accounts of explanation developed in the philosophy of science during the 20th century. Then they discuss the interrelations between explanation, understanding, grasping, and knowing (5:26). Next, they discuss Trout’s views on the role of empathy in explanation and understanding (16:35), evolutionary psychology in social science (21:37), and “the trap of the sense of understanding” (24:30). Then they turn to the practical benefits of explanation (37:11) and Strevens’s work on the level of detail in a good explanation (44:18). They conclude by discussing explanation outside of science (51:50).
Related works
by Strevens:
“No Understanding without Explanation” (2013)
“The Explanatory Role of Irreducible Properties” (2012)
Depth: An Account of Scientific Explanation (2008)
by Trout:
The Empathy Gap (2009)
“The Psychology of Scientific Explanation” (2007)
“Scientific Explanation and the Sense of Understanding” (2002)
[display_podcast]
Understanding: L- Routines vs. Non-L Routines:
1. For agents x and y there is a non-language (non-representational) routine A such that either agent has demonstrably more competency at than the other. There is some phenomena e that is operand in A.
2. For agents x and y there is a language (representational) routine L, such that either agent has demonstrably more competency at than the other. L purports to be about A; hence L purports to be about e.
(By ‘language routine’ I mean generally propositional though and it’s communication- inscription/enunciation: writing, scrawling, blogging, speech, discussion, debate, lecturing, etc. By ‘operand’ I mean e is included in the execution of A or A is an interaction between agent and e or there are constituent actions in A that are operant or respondent with respect to e etc.)
Question:
If x is more competent at L than y should this mean that x has more agency with respect to e?
At the Cider Saloon:
I enter a cider bar and sit next to a bearded gentlemen resembling Dan Dennett is every way. A few pints later our conversation turns to the hot topic of table tennis. The man’s eloquence, perspicuity and seamless comprehension leave my humble words for dead on the matter. Am I to assume that Dan’s doppelganger is better than I at the game? Is his eloquence a fair index of his otherwise situated and enactive sportsmanship?
Alternatively, is there a non-L routine that Dopple-Dan can execute involving the phenomena of consciousness say, such that he is clearly more competent at or perhaps, that I have no competency at all? What would this non-L routine consist in? To what extent is 2 really a guarantee of 1? And, if one’s general competency with respect to any e is exhausted by language routines, then what are we to say honestly regarding one’s agency with respect to e? Clearly x can have more armchair knowledge purporting to be about e than y, but both can be equally competent or incompetent in non-L routines where e is genuinely operand. Perhaps I was hoping in some opaque and under-qualified way to state that “I know as much as Dan Dennett does about consciousness” without that statement being completely false! Let’s assume that Dopple-Dan is in fact David Chalmers- pace inebriation- but this is irrelevant to the matter.
(Note that there are classes of non-L routines that depend heavily on the appropriation or deployment of L routines vis. studying to be a G.P, specialist, surgeon, petro/electrical/structural engineer, dentist or veterinarian say, will involve vast amounts of language activities before anything practical could be mastered. I also take it that L routines aren’t operant upon the things or phenomena that are supposed to have aboutness with respect to, perhaps with the exception of an agent capable of interpretation. I suppose yelling at the dog is operant upon the dog, but talking to my friend about my dog is not.)
Enactive Mary: Why L-routines by themselves won’t raise your agency.
Let’s say we highjack Frank Jackson’s ‘The Knowledge Argument’, but instead of having Mary learn everything there is about colour perception/qualia and knockdown arguments against physicalism. We have her learn everything science can teach regarding ‘surfing’: Studied under M. Turvey, Has developed an accurate dynamical systems model, knows all about the human movement aspects, kinematics, the physics of wave formation, the cultural history, associated sports (skate/snow/wake-boarding, the lingo, etc.). What could we expect of her ability to carve up a wave upon her release? Given that she was holed up in a room for such a lengthy period, would she even have the requisite fitness to get to get past the shore break- can Mary even swim? Let’s ad to this a girl called Jane. Jane could care less about science, but is an adept skate and snowboarder, but like Mary has never tried surfing. If we were to pit the two together is a surf comp, who do we think will fare better?
L-decoupling:
In L-routines purporting to be about e- propositional thought and its expression (speech and writing) – one can think, say and write whatever one wants. The only consequences if there are any are merely social ones. However when one interacts (enactively) with e one cannot do whatever one wants. I can tell you a story about how I sailed a cardboard box to the Mariana Trench, jumped into the ocean, swam to the bottom and had a game of chess with an angler fish, or I could tell you how I eat fire for breakfast. In these cases, apart from your disdain the world remains the same. If however I were to actually try these things the consequences would be obviously dire.
As actions L-routines are generally not operant or respondent with respect to the (principle) extension(s) of their subject matter (intensions s1 through sn). In cases of questioning or commanding persons or yelling at the dog to “get off the couch” L-routines remain essentially decoupled (not operationally coupled). Clearly thought and its communication is a kind of coupling on its own, but that system as a whole can only ever be coupled to its extension(s) via routines that are ultimately not linguistic; ones that consist in enactive operations -operant or respondent activities.
Why my 2 year old daughter needs to know that water is H2O seems irrelevant to any ‘operations’ she performs on or with water- she does understand it to an extent. In 20 odd years when she becomes a chemist, this situation will change. My beliefs about the constituents of bread remain ‘recreational’ until such a time that I actually try to make bread; in which case there would be something that I do- operant activity- where ‘yeast’ for example, is genuinely operand. Knowing that ‘water is H2O’ is redundant unless there is at least one (scientific) routine that is informed by it, e.g. performing an electrolysis experiment/snowflake formation.