Engineering doctrines
Cappelen and Plunkett open their 2018 paper “A Guided Tour of Conceptual Engineering and Ethics,” with a quote from Nietzsche’s Will to Power:
Philosophers … have trusted in concepts as completely as they have mistrusted the senses: they have not stopped to consider that concepts and words are our inheritance from ages in which thinking was very modest and unclear. … What dawns on philosophers last of all: they must no longer accept concepts as a gift, nor merely purify and polish them, but first make and create them, present them and make them convincing. Hitherto one has generally trusted one’s concepts as if they were a wonderful dowry from some sort of wonderland: but they are, after all, the inheritance from our most remote, most foolish as well as most intelligent ancestors. …What is needed above all is an absolute skepticism toward all inherited concepts.
It’s a good reminder that this diseasing of 20th century philosophy—its mixture of “functional monism” (the conceit of a one-to-one relationship between handles and reality) and “classical concepts” (as having sufficient and necessary conditions)—was not a universal affliction. It appears most of the century’s “great” thinkers grasped it as a problem far ahead of broader adoption or articulation. Dewey’s pragmatism is a de facto solution to verbal and frame dispute and thus an acknowledgment of terms’ untethering from wordly structures; Wittgenstein anticipates cognitive science by half a century; and Nietzsche writes in Twilight of Idols of a man “stuck in a cage, imprisoned among all sorts of terrible concepts.” Carnapian explication is undergoing a 180-degree revision, no longer considered defeated by Quine as previously assumed.
One enticing part of the Nietzsche quote, in my mind, is its comparison to the “mistrusted” senses. It would be egregious had philosophy not learned until recently to distinguish between the world as it is and the world as it is apprehended by our senses; I’d like to suggest that philosofolly’s previous conception of language and concepts is equally egregious.
There are so many causes of messy handles that even a cursory examination of the process by which humans invent and disseminate them subverts any modeling of them as “tidy” or monistic. But I’d like to introduce one frame, adapted from NNT: the Procrustean bed, where bad carvings—simplifications or Platonifications—amputate important parts of the whole, or else “stretch” it to fit a desired factoring. (Procrustes is a mythical figure who would lop off or stretch out his guests’ limbs in his home, so they would perfectly fit his bed.) It’s a barbarism masquerading as hospitality, so to speak. Think of the game of telephone, or Chinese whispers, by which each successive adapter unconsciously and consciously “stretches” or “amputates” the term as learned to fit with his models and experience of reality.
But now we have these garbage bins of concepts, and Cappelen & Plunkett want to improve them—engineer them. Construction has a rich vocabulary here that might diversify “engineering”: retrofit, renovate, refurbish, remodel, all of which have meaningfully distinct denotations. Perhaps someone will adopt them, but for now, the core questions of conceptual engineering, to C&P:
- How ought we assess concepts, both in determining which existing concepts need engineering, and, once engineered, whether the new formulations are improvements over predecessors?
- How important is the genealogy, or history, of a concept in making engineering decisions?
Expanding on the first question, what sorts of defects might our representational devices (“handles”) possess? The pair present:
- Cognitive defects (which distort reasoning)
- Moral or political defects (which undermine our values)
- Theoretical defects (which undermine philosophical progress)
- Semantic defects (incoherence or incompleteness)
While the entailments of #2 and #4 are very clear to me, the odd-numbered items are less so. I could suggest that the “all under one roof” problem of conflation and linguistic over-burdening is one of the central failure sources of conceptual disputes, but perhaps this is a “cognitive” or “theoretical” defect. Similarly, I would point to back-coherence (in the sense of preserving genealogy) as holding back concepts from being their best selves; still, I can’t pretend that maintaining the coherence of historical texts via such a historical continuity is unimportant.
(Sidenote: One core divide C&P attempt to grapple with is externalism or internalism, but until presented with a compelling argument otherwise, I’ll consider, in the spirit of “generalized compatibilism,” these stances as perfectly congenial in their abstract formulations. Internalism, C&P write, is the view that a concept or utterance’s “meaning depends on facts about” what’s inside an individuals “head”; externalism holds that “meaning is determined at least in part by facts having to do with the history of linguistic usage, or complex use patterns over time, or the judgments of experts.” I’ll classify this as a verbal dispute similar to literary theory’s Meaning Wars: it is clearly the case that the models in people’s “heads” [and bodies] are built from external norms and histories, and that there is both a descriptive fact about the society’s aggregate usage and the individual’s neural interpretation, and that the only source of contention at play is which of these ought to be called a concept’s “meaning.” Crucially, if two people or groups in contention don’t refer to the same “thing” in thingspace, they’re not having a disagreement; they’re talking past each other. Social meaning and individual meaning are clearly different “things”—or more accurately, “processes”—both clearly real.)
But I’m getting off-track: I want to talk about two doctrines I’d propose as crucial for any would-be conceptual engineer. There are plenty of smaller considerations which the current literature neglects; one is the metaphorical fabric which holds our concepts together and forms an important axis of meaning; the other is the linguistic conquests constantly performed on language, and the way these refactorings create variability in the field of meaning. But these addendums can come later; for now:
- The doctrine of non-intervention. Concepts should—in part because they can only, with any efficacy—be engineered locally. Only locals know the affordances of their specific use cases. Philosophers ought to engineer philosophical concepts, but leave fishing concepts to the fishermen. Engineering-on-behalf ought only provide possibilities for bottom-up adoption; it should never limit possibilities by top-down imposition.
- The illumination doctrine. Concepts should help illuminate the world, but never obscure it. This is especially important in ameliorative or ethical-political projects.
The non-intervention doctrine is partly an attempt to prevent the failures of “high modernism,” or “systemism,” or “overhaulism” (James Scott’s, John Gall’s, and Ari Holtzman’s terms, respectively). In areas of vast complexity, such as the intersection of language and human psychology, unknown nth-order effects dominate the landscape of consequences. Most recently, this hubristic approach was found in the behavioral economics of Kahneman and Tversky, who made grand, rationalistic claims about local domains like profession basketball, of which they knew absolutely nothing, naively supposing that the local expertise and folklore of these domains were merely riddled by cognitive biases (as opposed to having been weeded out based on what “works”—e.g. the hot hand “fallacy”). Perhaps the surest safeguard against the failures of overhaulism is allowing local communities to freely adopt those terms which work for them, and freely ignore those which don’t. To the would-be conceptual engineer I say: model and advocate, but never impose.
The illumination doctrine is founded on an ethics of accuracy. There is much danger in normative, ethics-based, “activistic” linguistic meddling. When top-down imposed, such conceptual engineering projects are well-known to us via writers like Orwell. While individual engineers almost without exception believe that their individual intervention on language is for the greater good (in the great tradition of noble lies), limiting the thinkable ought to be viewed as assailing free speech. (Recall also that once a technology is utilized for one political end, a tacit allowance of a class of strategy has taken place, and the technology will be co-opted by the other side: victimhood and identity politics’ adoption by the right is a recent & relevant example.) Expressive liberties are premised not just on the ability to say, but on the existence of an expressive language which enable the sayable.
Both these doctrines join the doctrine of preserved ground discussed in a previous post on conceptual engineering. There I distinguished between two types of linguistic “conquests” that re-carve the language, seeing the “divide and conquer” method as healthier for the language, in the long run, than the “narrow and conquer” approach:
First, the narrow and conquer method, where a specific sub-sense of a concept is taken to be its “true” or “essential” meaning, the core which defines its “concept-ness.” To give an example from discourse, Taleb defines the concept rationality as “What survives, period.” The second style I termed divide and conquer, where multiple sub-senses are distinguished and named in an attempt to preserve all relevant sub-senses while also gaining the ability to talk about one specific sub-sense. To give an example from discourse, Yudkowsky separates “rationality” into “epistemic rationality”—the pursuit of increasingly predictive models which are “true” in a loose correspondence sense—and “instrumental rationality”—the pursuit of models which lead to in-the-world flourishing, e.g. via adaptive self-deception or magical thinking—rationality as “what works.”
Conquests by narrowing throw out all the richly bundled senses of a concept while keeping only the immediately useful—it’s wasteful in its parsimony. It leaves not even a ghost of these other senses’ past, advertising itself as the original bundled whole while erasing the richness which once existed there. It leads to verbal disputes, term confusion, talking past each other. It impoverishes our language.
Division preserves the original, bundled concept in full, documenting and preserving the different senses rather than purging all but the one. It advertises this history; intended meaning, received meaning—the qualifier indicates that these are hypernyms of “meaning,” which encompasses them both. Not just this, but the qualifier indicates the character of the subsense in a way that a narrowed umbrella original never will. Our understanding of the original has been improved even as our instrumental ability to wield its subsenses grows. Instead of stranding itself from discourse at large, the divided term has clarified discourse at large.



