If I believe in elves that can be seen, then I do not hold that the visual appearance of elves is necessarily either an illusion or an hallucination. If I don’t believe in the moon, then I hold that the visual appearance of the moon is necessarily either an illusion (the moon is really something else with wrong apparent properties) or an hallucination (there is nothing there). What you believe is real partly determines your beliefs that how things seem is how things are, i.e., is accurate.
When some experience is an illusion or hallucination, it is in error in some way. As I’ve argued in an earlier post, how it is in error depends (at least in part) on how things appear (a phenomenological condition) and how things are really (a metaphysical condition). What one holds about the latter condition (at least) is theory-dependent. Thus, that there is an error — and what kind — is also theory-dependent.
The error can be different according to different theories even if: (a) everyone agrees as to what is really going on; they need only disagree about what appears to be going on; (b) everyone agrees as to what seems to be going on; they can disagree about what is really going on.
From this, I think that what is erroneous can depend on what one holds to be real about the senses themselves — and the properties we ascribe to them.
Turning this to errors in the direction of gaze (or visual direction), I think that there are three factors to consider:
(a) The definition of the direction of gaze
(b) The possibility of experiential error of direction which it neither illusion nor hallucination (anosognosia).
(c) The possibility of an experiential error of direction at all
Here is my reasoning for taking (c) seriously.
First, as argued elsewhere on the site, I don’t presume the universality of illusory counterparts. Just because something can be apparent doesn’t mean that it can actually be illusory. In the widest set of possible worlds, yes – it’s possible that what’s apparent is merely illusory — but it is an open question with respect to the actual world.
I think that the greater share of claims to error of direction are discrepancies between
(i) The apparent direction of something and
(ii) Its direction relative to a Earth-derived geometric system idealised as a sphere.
However, I also think that there is no particular reason to evaluate the accuracy of (i) by (ii). If one is to evaluate it by anything, it should be
(iii) A direction relative the path along which light travels from the source to the eye.
However, (i) is always accurate when evaluated with respect to (iii) light. That is, if something appears to be straight in front of you in mirages, loomings, or other strange visual phenomena, this is not, strictly speaking, erroneous. By the standard of the path of light itself, it is straight in front of you. This is the case if you choose the path of light from the source to your eye, the very thing that allows you to see the source. If you choose that path, then the source is straight in front of you and even orientated as it appears to be; that apparent direction and orientation is not mistaken.
(You might choose some other path of light to evaluate it, and judge it wrong. Fine — but why? One path of light is as good as another. If you are arbitrarily doing so, then the error here is at best in the absoluteness of apparent direction and orientation (and distance — but I’m ignoring that in this post)).
I think that will sound very wrong to many readers. However, I think, when it comes to considering the alternatives — the body, the world, the head etc. etc. — light is still the better option. All other options lead to cases of error that are intuitively not error. This option leads to cases of veridicality that are intuitively error. Since these cases are also cases of perceptually apparent veridicality (simply by being perceptually apparent) then we have a clash between (a) appearances of veridicality and (b) intuitions of error. In such a clash, where neither side is incoherent, I pick (a) every time.
Let’s look at some alternative ways of picking out the correct direction to judge experienced direction.