Comments on Gualtiero Piccinini’s ‘First-Person Data, Publicity, and Self-Measurement’

 

1. Introduction

 

GP’s paper has two main goals: 

 

The first is negative:  To show that privatism is false.  By ‘privatism’ GP means (briefly put) the view that  “first-person data are private but nevertheless scientifically legitimate”. 

 

The second goal is positive: To develop a novel account of first person data.  Let’s call the view that GP proposes the self-measuring instrument view.  GP says that on this view it’s the case that first-person data are both public and scientifically legitimate.  As he puts it, on the self-measuring instrument view “first-person data are as public as any other scientific data: their use in science is legitimate, in accordance with standard scientific methodology.”

 

I’ll split my comments between talking about some of the work that GP does to try to achieve his first goal and talking about some of the work that GP does to try to achieve his second goal.  In both cases I’ll try to make a close reading of some short but important sections of GP’s paper.  The trade-off here is that there are a lot of the interesting talking points in the paper that I won’t even touch on.  But hopefully ES’s comments and the conference comment thread will get to many of those. 

 

2. GP on privatism

 

After GP’s initial argument against privatism, he takes on an important response that many privatists would be apt to make.  He writes: “[A] common defense of first-person data by privatists is that although the deliverances of introspection are private, they can still be used as scientific evidence because introspection is reliable.  Or at least, it [introspection] should be taken to be reliable in the absence of evidence to the contrary”.  GP cites Chalmers and Goldman as amongst the privatists who try this line of argument. 

 

Let’s look at GP’s response to this defence of privatism.  I’ll try to reconstruct his response in a more standardized form.  Hopefully this will help us narrow in on individual parts of his response in an efficient manner.  As I read it the main thrust of GP’s response is best reconstructed as proceeding in two steps. 

 

GP’s Response to A Common Defense of Privatism: STEP#1

P1: If the deliverances of introspection are private, no independent evidence can be brought to bear on the reliability of introspection.

P2: It can be publicly established whether introspection is reliable only if we can bring independent evidence to bear on the issue.

C1: If the deliverances of introspection are private, then it can’t be publicly established whether introspection is reliable. (P1, P2 HS)

 

I take that P1 and P2 here are supposed to basically be definitional.  P1’s consequent is supposed to follow from its antecedent once we get settled on an adequate understanding of what it is for evidence to be private.  Similarly P2’s consequent is supposed to follow from its antecedent once we get settled on an adequate understanding of what it is for evidence to be public. 

 

This all seems ok to me.  I have no problem with P1, P2, or C1.  But it seems clear that C1 isn’t a conclusion that in and of itself is trouble for privatism.  Privatists can accept C1.  They can even say – as they probably should – that C1 simply states an analytic entailment of holding that the deliverances of introspection are private. 

 

We need to look at the second step of GP’s response to get to some conclusions that are in and of themselves trouble for privatism.  I take it that C1 is supposed to be an intermediate conclusion on the way to these troublesome conclusions.

 

GP’s Response to A Common Defense of Privatism: STEP#2

C1

P3: If it can’t be publicly established that introspection is reliable, nothing warrants the suggestion that introspection is reliable to any degree.

C2: If the deliverances of introspection are private, nothing warrants the suggestion that introspection is reliable. (C1, P3 HS)

C3: So privatists are, contra claims of some privatists, in no position to hold that “although the deliverances of introspection are private, they can still be used as scientific evidence because introspection is reliable”. (From C2)

 

I think STEP#2 is problematic.  I suspect P3 is false.  I’ll raise two responses to P3 that GP doesn’t consider.  Minimally I hope to show that GP needs to say more to justify our believing P3.  Also, though I hope it’s not the case that I’ve misrepresented GP’s response above, I think I’ve got it right that P3 is a central plank of his response.  Among other things, P3 is, except for some trivial modifications, a direct quote lifted from GP’s discussion of the issue at hand.

 

1st response to P3: 

One response to P3 seems to be the following.  Even if it can’t be publicly established that introspection is reliable, we could still hope to make a rigorous inference to the best explanation to justify belief that introspection is reliable.  We could, in short, be both privatists about introspection and explanationists about justifying the reliability of introspection. 

 

I’ll try to show roughly how this response could go by making an analogy with appeals to explanationism in theoretical physics.  Here are four interesting things Leplin says in his A Novel Defense of Scientific Realism about some contemporary work in theoretical physics:

 

1) That “[t]heoretical physicists, at least those working in elementary particle theory or, more specifically, quantum gravity, are unifiers.  They have doubts about superstring theory and particular GUT symmetries, but they do not doubt that the forces of nature instantiate some unifying symmetry.  They do not doubt that gravity can be quantized and unified with the other forces.  They believe that matter is unstable, and that under conditions that prevailed in the early universe elementary particles lose their distinct identifies”.

2) That physicists believe the things listed in 1) (e.g., are unifiers) partly because – or at least act as though they believe the things listed in 1) partly because – they subscribe to the following principle: The evidence that it is appropriate to require for the acceptance of a new scientific theory is the strongest evidence that can forseeably be obtained on the assumption that the theory is correct.  As long as the new theory fits the prevailing standards of theorizing – it has an appropriate mathematical formulation; [it isn’t self-certifying;] it addresses problems and promotes desiderata that established theories identify as important; it is consistent with the established body of scientific knowledge or, where inconsistent, rationalizes its departures in ways that prevailing theories recognize as legitimate – it is not to be ruled inadmissible for failure to be amenable to existing standards of empirical evaluation. …  [In short the view is that] insistence on conditions that the theory could not satisfy, whether or not it is true, is simply too restrictive.  Standards of evaluation are responsive to the content of the theories required to meet them.” (emphasis in original)

3) That with the standard of justification given in 2) physicists “are resorting to a standard of justification that they would much prefer to be at liberty to reject”.

4) That new unifying theories in physics are, consistent with the standard of justification in 2), evaluated based (very roughly put) on their explanatory power rather than on testable results. 

 

If we assume these 4 points are true, then we can quickly identify a way for privatists to try to argue that introspection is reliable that doesn’t depend on an appeal to public evidence.  The route parallels how unifiers in theoretical physics attempt to argue, for example, that gravity can be unified with other forces without producing testable results in support of that view. 

 

The route is for both privatists and unifiers is to develop rigorous formulations of their views, deliver evidence that satisfies the bar in 2), and try to show the explanatory power of their views.  The ‘deliver evidence that satisfies the bar in 2)’ clause is obviously key here.  Given the standard of justification in 2), it’s simply inappropriate to demand that privatists supply public data to justify their view that introspective data are both private and reliable.  The reason is that such data is in principle unobtainable if, as privatists claim, introspective data is private.

 

So this 1st response to P3 comes to the following head.  A rigorous explanationism of the sort employed in theoretical physics may be a potential method of justification for privatists to appeal to.  It offers a set of scientifically credible evidentiary standards – indeed evidentiary standards actually used in hard science – that privatists can argue are appropriate to apply when evaluating their hypothesis that introspection is private and reliable.  Similarly privatists can argue that if they can meet these standards then, contra P3, their hypothesis that introspection is private and reliable can be warranted even if it can’t be publicly established that introspection is reliable. 

 

I think the sort of appeal to explanationism sketched here is well worth considering.  That said I have to admit that if I were a privatist the prospect of appealing to explanationism would make we feel rather nervous.  The same nervousness seems to felt amongst theoretical physicists – as 3) indicates.  So the 2nd response to P3 that I’ll offer may seem preferable to many privatists.

 

2nd response to P3:

A second response to P3 seems to be the following.  Even if it can’t be publicly established that introspection is reliable, there may be good epistemological reasons to believe that introspection is a reliable evidentiary resource.  I have in mind a reason that Goldman has recently mentioned in some of his work of intuition (“Philosophical Intuitions: Their Target, Their Source, and Their Epistemic Status” Grazer Philosophische Studien, 2007).  But the argument has legs (going back through Alston and others) in arguments for and work on foundationalist theories of the structure of epistemic justification.

 

Goldman puts the argument as follows:

 

“There must be some procedures or methods [of acquiring evidence] that are basic. In other terminology, there must be some basic “sources” of evidence.  Basic sources are likely to include mental faculties such as perception, memory, introspection, deductive reasoning, and inductive reasoning.  These faculties are all regarded, by many or most epistemologists, as bona fide sources of evidence.  Yet all or many of these sources may be basic in precisely the sense that we have no independent faculty or method by which to establish their reliability.  Yet that doesn’t undercut their evidence-conferring power. Consider memory, for example.  Memory may be our basic way of forming true beliefs about the past.  All other ways of gaining access to the past depend on memory, so they cannot provide independent ways of establishing memory’s reliability [Goldman references Alston’s The Reliability of Sense Perception here].  If we accept [the constraint that x is valid evidence only if x has independent corroboration] the outputs of memory will not constitute legitimate data or pieces of evidence.  But that is unacceptable, on pain of general skepticism.  It is better to accept the conclusion that basic sources of evidence don’t have to satisfy the … independent corroboration constraint.” (emphasis in original)

 

Although Goldman is focusing on the example of memory here, it seems there’s an analogous argument regarding each evidentiary source that foundationalists will call ‘basic’ sources of beliefs – and, so, for introspection.  The argument is roughly the following:

 

The Basic Beliefs and Independent Evidence Argument:

P1: All ways of gaining access to the mind ultimately depend on introspection.

C1: So, no ways of accessing the mind can provide introspection-independent ways of establishing introspection’s reliability.  (From P1)

C2: So, if we endorse an independent evidence criterion for x being valid evidence, the outputs of introspection won’t count as valid evidence. (From P1)

P2: But C2 is unacceptable, on pain of general skepticism about the mental.

P3: A better route is to accept that some sources of evidence – call them, as foundationalists are apt to, ‘basic’ sources of evidence, e.g., introspection – don’t have to satisfy an independence evidence criterion in order to accepted as reliable sources of evidence.

C3: Even without independent evidence in favour of its reliability, we can endorse it being the case that introspection is reliable.  (From P4, P5)

 

This argument has three substantial premises.  I won’t try to defend them.  Also it will also be obvious that I haven’t formulated the three substantial premises too rigorously.  So tighter versions of the argument can no doubt be generated. 

 

But the main point is that the general bent of the Basic Beliefs and Independent Evidence Argument seems to raise an important worry for P3 in GP’s Response to A Common Defense of Privatism.  It suggests that work on foundationalist theories of the structure of epistemic justification offers privatists two things.  First, plausible grounds to reject P3 in GP’s Response to A Common Defense of Privatism.  Second, plausible grounds to argue that we should believe that introspection is a reliable evidentiary resource.   

 

3. GP’s self-measuring instrument view

 

I find GP’s self-measuring instrument view pretty intriguing.  I want to focus in on one particular part of the view. 

 

An important feature of the self-measuring instrument view is the place of training or calibrating subjects.  The basic idea is that, roughly on analogy with more ordinary measurement devices (e.g., thermometers), subjects generally need to be trained or calibrated so that they yield valid and useful first-person data. 

 

Regarding calibration, I read GP as making the following argument:

 

GP’s Calibration Argument

P1: We want to obtain valid first-person data in certain psychological experiments

P2: If we’re to obtain valid first-person data, the effects of such confounding factors as illusion, delusion, forgetfulness, confabulation, wishful thinking, lying, etc. amongst test subjects must be minimized.

P3: If the effects of such confounding factors must be minimized, we can hope to do this sometimes via calibrating subjects. 

C: We can hope to obtain valid first-person data sometimes via calibrating test subjects.  (P1, P2, P3 MP MP) 

 

Both P1 and P2 seem very plausible, to say the least.  So the weight here is on P3 – as it presumably should be since the main thrust of the argument concerns the usefulness of calibration.  So let us turn to GP’s argument for this P3. 

 

Here is a reconstruction of GP’s argument for P3 in his Calibration Argument:

 

GP’s Children and Teenagers Argument

P1: In some training / calibration cases, e.g., certain child rearing and teenagers calling each other on their bullshit cases, we can bring independent evidence to bear on the issue of whether a subject is exhibiting a confounding factor.

P2: If P1 is true, then if the effects of confounding factors must be minimized, then we can sometimes hope to do this via calibrating subjects.

C: If the effects of confounding factors must be minimized, then we can sometimes hope to do this via calibrating subjects.  (P1, P2 MP)

 

This reconstruction is a bit inelegant.  But it will do for our purposes.  Howsoever GP’s argument for P3 of his Calibration Argument is reconstructed, something at least very similar to P1 of what I’m calling GP’s Children and Teenagers Argument will figure prominently.  And it’s this premise that I want to address.  So let’s turn to addressing it. 

 

I think there are troubles both with the case that GP makes for P1 in his Children and Teenagers Argument and, probably, with P1 itself.  In what follows I’ll try to make one point on each of these fronts. 

 

1st point

It seems to me that there is a gap in the support that GP provides for P1 in his Children and Teenagers Argument when he sketches his child rearing and teenage bullshit cases.  The gap is to tell us what exactly the independent evidence is in these cases that is purportedly being brought to bear to evaluate the relevant person’s first-person data for confounders. 

 

One response to this request to fill-in this gap might be to say that the particular evidence doesn’t need to be pinpointed.  The idea would be that what matters is just that the child rearing and teenage bullshit cases (and the other cases GP offers in the same vein) show, as GP says, that “we have plentiful ways to tell when others are inaccurate about their mind”. 

 

But this response isn’t adequate.   Even if it’s the case that we have plentiful ways to tell when others are inaccurate about their minds, it doesn’t immediately follow that some of these ways involve bringing independent evidence to bear on whether others are inaccurate about their minds.   Our ways of telling might proceed exclusively along lines that don’t use independent evidence.  For example, they might just be inferences to the best explanation that are based on non-independent evidence.

 

2nd point:

Part of the trouble with GP not pinpointing the independent evidence at work in the child rearing and teenage bullshit cases is that it’s not easy to fill in the blanks for him here.  For time reasons let’s stick to just working through one teenage bullshit case to see this.   The moral we draw should generalize well to the other sorts of cases that GP mentions.

 

GP doesn’t develop a particular case of teenagers calling each other on their bullshit.  But I take it he has something like the following in mind:

 

Suppose Amanda says something like ‘I don’t care that he dumped me for Jessica.  Screw him.  Jessica’s a tramp and he’ll just get what he deserves anyway.’  If Amanda’s friend Florence hears this and is trying to simply be supportive, she might just say ‘Yeah, totally’ in response.  But if Florence wants to make a more penetrating response, she might say ‘Amanda, you’re just bullshitting yourself if you think that you don’t care that he dumped you.  If you didn’t care, you wouldn’t be lacing into him by saying ‘Screw him’, or calling Jessica a tramp, or saying he’ll just get what he deserves’.   

 

This looks like a potential case of one teenager trying to call another on her bullshit.  (Well, assume Amanda and Florence are teenagers.)  Further it has at least some potential to be a case where Florence has independent evidence for calling Amanda on her bullshit.  The potential independent evidence is the data of the second half of Amanda’s initial utterance – that is, the part where Amanda says ‘Screw him.  Jessica’s a tramp and he’ll just get what he deserves anyway’. 

 

The question for us to try to answer, then, is whether the second half of Amanda’s utterance is independent evidence to bring to bear on the reliability of the first half of Amanda’s utterance. 

 

I think the answer is, minimally, that it’s not clear that it is.  The second half Amanda’s utterance seemingly can’t be independent evidence against the reliability of the first half of Amanda’s utterance because these two pieces of first person data are both outputs of the same procedure.  The procedure here is the introspection-based procedure of moving from thought to action (where the action in Amanda’s case is a verbal utterance).  

 

It’s worth mentioning three further things about this teenage bullshit case:

 

First:  In the case I described I said very little about Florence.  She hears Amanda’s utterance.  Then she provides a response.  Florence seems to very much be in the situation that a scientist recording Amanda’s or someone else’s first person-data would be in.  That is it seems that basically the same first person data is available to Florence as to the scientist.  In short then it seems that the main problems of independent evidence generalize from the Amanda-Florence case to subject-scientist cases. 

 

Second:  GP says that one asset of his self-measuring instrument view is that on it “[t]he epistemic burden of establishing how accurate, reliable, and precise the data are falls on the scientists, not the subjects”.  But given what we’ve just seen, it’s not clear how advantageous this trait is.  The trouble of getting independent evidence to assess Amanda’s first person data from the standpoint of Florence or the scientist seems great. 

 

Third:  It may bear noting as well that the problem here isn’t alleviated if Amanda says or does a whole lot more.  All the first person data we can gather from Amanda’s first person behaviours seems to, at least to some measure, be an output of introspection – with the exception of reflexes and other non-intentional behaviours.  Each bit of data from Amanda’s stream of intentional first person behaviours seems to be non-independent of each other bit.

 

So this 2nd point comes to the following two heads.  On the one hand, it’s not clear that we should grant that what’s going on in everyday cases like child rearing and teenagers calling each other on their bullshit amounts to calibration – at least so long as we require, as GP seems to and as is the norm in science, that calibration requires the bringing of independent evidence.  On the other hand, it’s a challenge to see where the independent evidence can come from to calibrate introspection even in experimental settings.  The self-measuring instrument view seems to depend on the idea of being able to reliably calibrate subjects.  So these calibration issues directly bear on the prospects of the self-measuring instrument view.