Friday, October 23, 2015

A Grand, Unified Theory of Moral Metaphysics, Psychology, and Semantics

Professor Tresan

[nb: undefended controversial assertions herein]

what’s a property?
moral realism
naturalistic moral realism
two challenges: coherence & explanation
why do we care about moral properties?
the enforcer’s perspective
the navigator’s perspective
the navigator’s perspective and the coherence of naturalistic moral realism

what’s a property?
synonyms or near synonyms: feature, characteristic, trait, attribute, quality, type
a property is a way of being; any way a thing can be or said to be is a property
when we describe things we say what their properties are
if we say “x is F”, what “F” picks out is a property
if the sentence is true (x really is F) then x instantiates the property picked out
(but properties are not linguistic things; words “pick out”, “express”, “mean”, or “ascribe” properties)

things have properties; fancier ways of saying this are that things instantiate or exemplify properties
or that the property is instantiated or exemplified by the thing
a thing which has a property is a bearer of a property (or an “instance” or “token”)
e.g., an apple which is red is the bearer of the property of redness
the set of all bearers of a property is sometimes called the “extension” of the property
e.g., the extension of redness is the set of all red things (red apples, blood, the flag of China, etc.)
properties aren’t identical to their bearers
properties are ways of being, the bearers are the things which are those ways
properties aren’t even identical to their extensions
[for one thing, distinct properties can have the same extensions;
for another, any property could have had a different extension; there could have been a different set of red things]
properties are had in common by their instances, and nothing else

properties are extremely abundant
and are had by many different sorts of things [here are just some examples]
physical objects have properties
e.g., shapes, colors, weights, locations, constituents, movements, histories, causes, effects, functions
[re the last: some physical objects have the property of being a chair, or table, toaster, car, shirt, shoe, pen, knife, money, etc.  these are functional properties in the sense that they involve doing something or being used in some way; they are thus relational and not intrinsic properties; e.g., what makes a bill money is not its material constitution but its social role; note that it makes sense that our standard terms for these things are terms for functional properties given that that’s what’s most salient and interesting to us when we talk about them.]
people have properties
e.g., physical properties, but also psychological ones
[e.g., having a visual experience, thought, feeling, memory, fantasy, expectation, intention, emotion, desire, etc.]
actions have properties
e.g., temporal location, bodily motion constituents, causes (e.g., intentions), effects, being of some behavioral type
[re the last: e.g., being walking, talking, dancing, repairing, violin-playing, reading, making/keeping/breaking a promise, telling a lie/the truth, being cruelty, treachery, hogging, shirking, maximizing happiness or failing to]
properties have properties; e.g., redness has being a color, being Jones’ favorite color, being had by blood, etc.

properties can be complex (composed of other properties);
e.g., being square = being a closed figure, having four sides, having equally long sides, having 90° internal angles
properties can be indeterminate (have borderline cases)
e.g., being rich, tall, nearby, a short story (how many pages?), a heap, jazz, old, mature, bluish-green, shouting
properties can be hard to articulate with any precision, yet still be recognizable (many examples already given)

facts and states of affairs are often taken to consist in things having properties
e.g., the fact that snow is white = snow’s having the property of being white.

moral realism
we speak as if there are moral properties
e.g., we say “killing is morally wrong”, which seems to mean that killing has the property of being morally wrong
if we wonder whether some action is right or wrong, we seem to be wondering about its moral properties
moral realists say that there are instantiated moral properties
e.g., moral goodness, badness, rightness, wrongness, permissibility virtuousness, viciousness
[for ease, i focus on wrongness, and omit the qualifier “moral”]
moral realism seems plausible: didn’t the nazis do things which instantiated wrongness?
don’t lots of your actions instantiate moral permissibility?  (e.g. what you’re doing right now)
for realism to be true it’s only necessary that one thing one time instantiated one moral property
so endorsing realism doesn’t require endorsing any particular set of moral views
if realism is true, the extension of, say, wrongness is the set of all wrong actions
so the bearers of wrongness are actions
people and character traits are also the bearers of moral properties (e.g., Jones is morally good, honesty is a virtue)
the relations between moral properties is interesting but not one i’ll pursue here

naturalistic moral realism
if there are moral properties, there’s a lot want to know about them (e.g., their extensions)
one thing we want to know if their metaphysical nature: what kind of properties are they? 
one way of bringing out one issue of this sort is via an argument against moral realism
the argument is inspired by a passage from Hume:

Take any action allow’d to be vicious: Wilful murder, for instance. Examine it in all lights, and see if you can find that matter of fact, or real existence, which you call vice [or wrongness]. In which-ever way you take it, you find only certain passions, motives, volitions and thoughts [and other psychological events, as well as physical events, behaviors, causes and effects]. There is no other matter of fact in the case.

let us call the underlined properties of actions “natural” properties
they can be the object of non-evaluative awareness (e.g., by witnesses, reporters, social scientists)
that is, one can, in principle, take an action to have those features and not thereby evaluate the action in any way
e.g., consider the shooting of reporter Alison Parker and photojournalist Adam Ward
natural features of that action include the physical behavior of the shooter, the psychological causes of those behaviors (e.g., his emotions, desires, thoughts, etc.), and the physical and psychological effects on the victims.
if you were assigned simply to describe what he did without yet evaluating it, you could mention those sorts of things and wouldn’t be accused of having failed to do the assignment.

the argument:
1.             If anything has wrongness then that action has wrongness.
2.             If that action has wrongness then it has some property in addition to its natural properties.
3.             That action doesn’t have any properties in addition to its natural properties. 
so,
4.             That action doesn’t have wrongness. [2,3]
so,
5.             Nothing has wrongness.  [1,4]

the argument is valid and we can take 1 for granted, which means what’s crucial is whether 2 and 3 are true.
why think 3 is true? 
well, why think a claim like 3 is ever true? 
e.g., are these magical beans?  do they have magical properties?  if we don’t think so, why not? 
we are aware of no such property, nor of anything which is best explained by such a property
[that doesn’t mean it’s not there, just that we’re not justified in thinking it is there]
as Hume says: “Examine it in all lights, and see if you can find [such a property]”
the same seems true of the action and its natural properties
potential reply: could it be that our sense that the action is morally wrong is best explained by such a property?
that partly depends on how well we can explain that sense without positing such a property
i think we can explain it pretty well
plenty of moral realists reject premise 3 but I prefer to reject premise 2
that makes me a “naturalist” moral realist: i take moral properties to be natural properties
which natural property is wrongness? 
that’s a matter requiring significant investigation and controversy
remember that properties can be complex, indeterminate, and hard to articulate with precision even if recognizable
i take it that’s true of wrongness: we can recognize it, identify many instances, note some general patterns,
but a precise articulation of what it is for an action to be wrong isn’t easy.
but to illustrate what naturalists are saying, one naturalist view is that wrongness = failing to maximize happiness
i.e., A is wrong = A is such that the amount of happiness in the world as a whole (spatio-temporally) is less than there would have been if the agent had done something else which was available to her at the time.
note that whether an action fails to maximize happiness can be the object of non-evaluative awareness.
e.g., it’s common ground that the action in Transplant maximizes happiness; no position on its morality is implied.

two challenges to naturalistic moral realism
i’m going to spend the rest of the presentation on two challenges to naturalistic moral realism.
the first is a challenge to its coherence; the second is a challenge which confronts every theorist.
the plan is to pursue the second challenge and thereby eventually find a reply to the first.

the first challenge: is naturalistic moral realism coherent? 
here’s an argument that it isn’t:
1.             If wrongness is a natural property then wrongness can be the object of non-evaluative awareness.
2.             If wrongness can be the object of non-evaluative awareness then we can take an action to have wrongness without thereby evaluating that action.
3.             Taking an action to have wrongness = taking it to be wrong. 
4.             If we take an action to be wrong we are thereby evaluating it. 
So,
5.             If we take an action to have wrongness we are thereby evaluating it.  [3,4]
So,
6.             Wrongness can’t be the object of non-evaluative awareness. [2,5]
So,
7.             Wrongness isn’t a natural property.  [1,6]

the argument is valid and 1 and 2 are true by definition, so everything depends on whether 3 and 4 are true.

the second challenge: explain why humans exhibit moral attitudes and practices

let’s consider some well-known, uncontroversial anthropological facts
[attitude = a way of caring about something, hedonically, emotively, or motivationally
any psychological state directed at something which is inconsistent with its owner being utterly indifferent to it counts as an attitude]

the central fact: moral judgments are correlated with co-targeted, co-valenced attitudes
e.g., judgments that x-ing is wrong are correlated with dislike of x-ing
the correlation holds even in the absence of independent concern about the effects of the target
e.g., people dislike actions they judge wrong, independently of their dislike of the effects of those actions
e.g., consider moral outrage at actions learned about on the internet, involving complete strangers
[wrinkle: many also care about the welfare of total strangers but that doesn’t fully account for the dislike of wrongdoing]
the correlation holds even in the absence of independent concern about the agent
e.g., when strangers act wrongly we dislike that, even if we have no other dislike for those agents
the intensity of the correlated attitudes tends to covary with the extent to which moral properties are taken to be exemplified
e.g., what’s seen as mildly wrong (stealing gum) is disliked less than what’s seen as very wrong (atrocities)
the terms used to express moral judgments are used to display, guide, and mobilize the correlated attitudes
e.g., everyone recognizes that “wrong” is used to display, guide, and mobilize attitudes like aversion, disapproval, and outrage.  [suppose people start tweeting and re-tweeting that what someone did was “wrong” – why might they be doing that?  not just why they take “wrong” to correctly apply to the action, what are they trying to achieve by publicly labeling it?]

the explanatory challenge is: why are we like this?  (the ways reported above) 
if moral realism is true then at one level the answer seems pretty easy: people care about moral properties, i.e., about whether and the extent to which they’re instantiated – that’s why we think and talk about them, and why that thinking and talking is related to attitudes in the ways reported.
[henceforth “care about properties” = care about whether and to what extent they’re instantiated]
e.g., people dislike the instantiation of wrongness
and they care about the properties as such, hence the independence from concern about the effects and agents
as such = not just derivatively of their concern about something else (e.g., we don’t care about money as such)

but that just raises a further question: why do we care about moral properties in those ways?
if naturalistic moral realism is true then moral properties are natural
e.g., wrongness is a natural property by, e.g., brutality, betrayal, hogging, and shirking
so the question for naturalistic moral realists is: why do we care about those natural properties in those ways?

why do we care about moral properties? (the enforcer’s perspective)

the core hypothesis: caring about moral properties as such efficiently, reliably promoted cooperation in the EEA
(EEA = environment of evolutionary adaptedness – the environment in which the relevant features of our psychology evolved – last few million years)
hunter-gatherer groups competed with each other, both directly (by war) and indirectly (e.g., resource extraction, child-rearing)
groups which cooperated better did better at this competition, thus increasing their impact on the overall genome
cooperation in them was promoted by various behaviors and undermined by various behaviors
e.g., undermined by brutality, betrayal, hogging, shirking
the extent to which those behaviors occurred was influencible by attitudes in groups
e.g., pro-attitudes to the cooperation-promoting behaviors would encourage them,
con-attitudes to the cooperation-undermining behaviors would discourage them.
caring about the behaviors as such was efficient in that no further cognitive resources were required for the relevant response to be elicited (e.g., treachery elicits con-attitudes directly, not just when subjects are aware of certain further features of it, such as its effects). 
it was reliable in that it would have fewer false negatives than other, similar attitudes (e.g., attitudes about behaviors with certain effects, or in or directed at one’s group) [made welcoming newcomers easier]
the lack of agent- or victim-restrictions might result in false positives; e.g., disliking mistreatment of those in other groups, even by members of those groups, which seems at best wasteful
but limitations on awareness in the EEA would tend to limit those false positives
as would nonmoral attitudes (e.g., egoism, nepotism, tribalism). 
we thus find the development of an “enforcer’s perspective”: a cluster of traits designed to encourage cooperation-promoting tendencies and discourage cooperation-undermining tendencies
moral properties are such that attitudes to them as such would efficiently and reliably have those effects.
practices of moral judgment and socio-linguistic practices both manifest those attitudes and are independently selected for (co-functional)

why do we care about moral properties? (the navigator’s perspective)
a further important explanandum: the differences in our self- and other-directed moral attitudes
e.g., re others’ wrongdoing: disapproval, outrage, indignation, etc.
re our own wrongdoing: guilt, shame
we are especially interested in ourselves acting well
for many, we care about this as such (not just to avoid punishment)
a hypothesis: these self-directed moral attitudes are due to the prominence of the enforcer’s perspective
once the enforcer’s perspective was prominent in the social environment, it constituted a selection pressure
our ancestors who acted in ways which elicited widespread con-attitudes did worse
those who acted in ways which elicted widespread pro-attitudes did better
thus, traits promoting the navigation of enforcers’ attitudes (moral norms) were selected for
this yielded a “navigator’s perspective”: a cluster of traits designed for efficient, reliable navigation of moral norms
self-directed moral attitudes efficiently and reliably promoted this navigation
so the proposal is that there was a two-stage process
first, the enforcer’s perspective emerged to promote cooperation
second, the navigator’s perspective emerged to promote navigation of enforcers (moral norms)
[note: this is compatible with our caring about the morality of our actions and character as such; the lust point]

the navigator’s perspective and the coherence of naturalistic moral realism
self-directed moral attitudes aren’t the only traits which help navigation of moral norms
there are also traits designed for the tracking of those norms (i.e., awareness of them)
roughly, navigating norms requires not just wanting to navigate them but knowing what they are
moral norms are a hugely significant feature of our environments, like predators, snakes, cliffs, enemies, etc.
e.g., which actions will elicit wrong-judgments is significant: if you don’t know what they are, you might perform them, thus drawing wrong-judgments with all the negative consequences of that.
we are thus designed to track wrong-judgments (which wrong-judgments those in our social environment are apt to make)
that is not essentially a moral motive; it is deeply practical
e.g., psychopaths who don’t care about moral properties as such still have a motive to care about moral norms
the crucial point: our typical states of mind when we’re aware of moral judgments is deeply different than our states of mind when making moral judgments
the former is a very practical perspective, designed to help us navigate our social environment
the latter is a perspective dominated by concern about moral properties as such
this hypothesis has certain linguistic concomitants
if what’s salient to us when moralizing are moral properties, it makes sense for us to use terms for those properties
and that’s what realists say about moral semantics; e.g., that “wrong” when used for moralizing means wrongness
if what’s salient to us when we’re aware of moral judgments is not the properties they’re about but what we do with them then it makes sense that our standard terms for moral judgments would pick them out in virtue of this role.
the suggestion, then, is that this is what we do:
when we say “x-ing is wrong” and thereby morally judge x-ing, we mean that x-ing has wrongness
when we say something like “they think x-ing is wrong”, we mean that they’re opposed to x-ing in a certain way, not that they think it has wrongness.
roughly, “judgment that x-ing is wrong” means judgment about x-ing which has the bolded relations to attitudes 

if so, then premise 3 of the incoherence argument is false. 
wrongness is the property we ascribe to actions when we judge that they’re wrong
so our judgments about what’s wrong are judgments about that property
but what makes them judgments that something is wrong isn’t that they’re about that property. 
it’s their role, their relations to attitudes and practices.
what makes 4 true is that when we judge an action to be wrong we’re thereby making a move in a certain attitude-involving practice
since merely ascribing wrongness isn’t making such a move, 3 is false.

a remaining issue: the (im)plausibility of the ambiguity claim
is it really plausible that “wrong” has a different meaning in “x-ing is wrong” and “judgment that x-ing is wrong”?
it goes without saying among metaethicists that this is not so.
but there’s a plausible explanation regarding why it might have occurred.
we start with what was suggested, that our perspective on moralizing is different than our perspective in moralizing
now we add a common process of semantic change: the conventionalization of token-for-type metonymies
metonymy involves the use of a term for x as if it is a term for some y with which x is saliently associated
it’s ubiquitous and has as many forms as there are ways of being saliently associated:
part for whole (“V8s are more expensive” for cars with V8 engines);
container for contained (“drank the bottle” for drank its contents);
location for institution (“White House announced” for administration announced);
order for customer (“the ham sandwich” for the customer who ordered it); material for item (“do you take plastic?” for credit cards);
token for type metonymies involve words for tokens being used for saliently associated types.
these form a broad class:
person for personality (“another Hitler” for another bloodthirsty megalomaniac);
brand for product (“rollerblades” for inline skates);
history for event (“no more Vietnams” for no more bloody quagmires) [etc.]
in the case at hand, the token is our judgments about wrongness, which in fact play the relevant role;
they type is judgments which play that role.
finally, conventionalization involves the transition from figurative to literal: a semantic change
and this is a wholescale occurrence, regarding moral terms in general (compare: “Mommy”, “Daddy”).