Friday 19 July 2019

Palmira’s Objection to Friedman

In his paper, “Inquiry and the Doxastic Attitudes”, Michele Palmira proposes a new “sui generis doxastic attitude which differs, both functionally and normatively, from suspended judgment, full belief, credences, and acceptance”(p. 1).  He calls this proposed attitude “hypothesis”.  I agree with Palmira that there is a need to posit a sui generis attitude with the normative and functional profile he ascribes to hypotheses.  However, I take issue with what Palmira claims to be one major upshot of the attitude’s existence.  Specifically, he takes the existence of the attitude of hypothesis to be inconsistent with the following thesis defended by Jane Friedman:

Necessarily: One is in an inquiring state of mind about some matter if, and only if, one is suspended about that matter.[1]

Palmira takes as his point of departure the observation that inquiry often involves three distinct stages: 
Three-stage model of inquiry 
  1. One is open-minded about how to answer the question Q. 
  2. One is inclined to answer Q in a given way while taking the question to be still open. 
  3. One closes Q. 
In both stages 1 and 2, a certain question, Q, is regarded as open by the inquiring agent. A question, Q, is open for some agent, S, just in case S does not believe or know that some proposition, P, constitutes a complete answer to Q.  What distinguishes stage 1 from stage 2 is that in the latter, the inquiring agent takes some particular answer to Q to be more likely to be true than all competing alternative answers.  Hence, while in stage 1, an agent may see two or three candidate answers to Q to be equally likely to be correct, given her available information, during stage 2, the agent comes to view one particular answer to be more likely to be correct and focuses her attention to gathering information relevant to the assessment of that particular answer. It is the act of zeroing in on a particular answer to a question one takes to be open for the purpose of engaging in more narrowly constrained investigation into said answer that Palmira identifies with the attitude of hypothesizing.  Moreover, by Palmira’s lights, the shift from stage 1 to stage 2 of inquiry is marked by a shift from suspending Q to hypothesizing PQ, where the expression “PQ” means that some proposition P is a complete answer to Q.[2]  In sum, Palmira insists that the distinction between stages 1 and 2 of inquiry is best captured by saying that while in stage 1 one has the attitude of suspending Q, in stage 2 one has the attitude of hypothesizing PQ.

This brings us to Palmira’s criticism of (BICON).  (BICON) predicts that the inquiring agent would be suspended at both stages of her inquiry.  This follows from the fact that (BICON) states that being in an inquiring frame of mind towards Q entails that one is suspended about Q.  Since both stages 1 and 2 involves being in an inquiring frame of mind towards a question, it follows from (BICON) that both stages 1 and 2 entail that one is suspended about that question.  Palmira puts the objection as follows:
Importantly, (BICON) contains an answer to the focal question of this paper: since being in an inquiring state of mind towards a given matter just is being suspended about it, and since one is in such an inquiring state of mind at both stages of open inquiry, it follows that one just is suspended at both stages of open inquiry. (p. 5)
Hence, according to Palmira, (BICON) does not leave room for the idea that stages 1 and 2 of inquiry implicate different doxastic attitudes: suspending and hypothesizing, respectively. This is Palmira’s first line of criticism.  His second line of criticism is that (BICON) runs into what he describes as the Unity of Inquiry Problem This second line of criticism begins with the observation that the inquiring agent is disposed to make different inquisitive moves during stages 1 and 2.  For example, suppose I am considering the following question:

Q1: Where did life on earth come from?

During stage 1, my investigation may focus on figuring out what are some of the candidate answers to Q1.  This initial period of investigation may culminate in my thinking that there are at least three equally likely answers to Q1:

P1Q1: Life on earth originated from non-life.
P2Q1: Life on earth came from some other planet.
P3Q1: Life on earth was created by an intelligent designer.

Suppose that after further investigation I come to regard P1Q1 as being the most likely answer to Q1.  At that point, I would have transitioned from stage 1 to stage 2 of my inquiry into Q1.  Plausibly, my investigation will no longer be concerned with figuring out what are the candidate answers to Q1, as was the case during the first stage of my inquiry into Q1.  Instead, my investigation will now be concerned with the evidence for and against life on earth originating from non-life. 

According to Palmira, the defender of (BICON) can only accommodate the above shift in the focus of my investigation by holding that the question I am suspended about has also shifted.  Whereas I was initially suspended about Q1, I am now suspended about something along the lines of Q2:

Q2: Did life on earth originate from non-life?

Hence, Palmira insists that (BICON) implies that at the different stages of inquiry the inquiring agent is suspended about different questions.  However, this would undermine the idea that the inquiry I am engaged in is a single unified activity.  Moreover, since (BICON) implies that I am inquiring about Q1 only if I am suspended about Q1, the fact that during the second stage of my inquiry I am suspended about Q2 rather than Q1 implies that I am no longer inquiring into Q1.  Here is the objection in Palmira’s own words:
[I]f we accepted the idea that one holds an attitude of suspended judgement at both the first and the second stage of the inquiry, we would not be able to capture the unity of subject-matter of the inquiry one is pursuing, for the content of the inquiring state of mind would change across stages and there is no other way to specify which [subject-matter of inquiry] is at stake. This is the Unity of Inquiry Problem. (Palmira, 2018: 6).
While I agree with Palmira that (BICON) should be rejected[3], I find his objections to the left-to-right side of (BICON) unconvincing.  Let us call the left-to-right side of (BICON)—i.e., the claim that being in an inquiring frame of mind entails suspension—the inquirer-as-suspender thesis.  Palmira’s first line of criticism is meant to show that the inquirer-as-suspender thesis does not leave room for the suggestion that stages 1 and 2 of inquiry implicates different doxastic attitudes; suspending Q1 and hypothesizing P1Q1, respectively.  However, this criticism only gets off the ground if we assume that being in an inquiring frame of mind towards Q at some time, t, entails that suspending Q is the only Q-related doxastic attitude one has at t.  However, this is not what the inquirer-as-suspender thesis claims.  What the inquirer-as-suspender thesis claims is that being in an inquiring frame of mind towards Q at t entails that suspending Q be among the set of doxastic attitudes one has at t.  Whether or not the set of attitudes in question is one that only has suspending Q as a member is a matter about which the inquirer-as-suspender thesis is entirely silent.

Moreover, the claim that suspending Q is the only Q-related attitude one may have when one is inquiring about Q is one Friedman explicitly rejects.  By her lights,  suspension accompanies (and is perhaps even a pre-requisite for) every other inquiring attitude.  Hence, Friedman is open to the possibility of an inquiring agent having other Q-related attitudes alongside suspending Q at a given point in time.  This means that if Palmira wishes to deny that an agent may rationally suspend Q at some time, t, and also hypothesizing PQ at time, he will need to provide us with some positive grounds for thinking that these two attitudes in particular are incompatible.

Palmira regards hypothesizing as a doxastic attitude.  It may therefore be argued that simultaneously suspending and hypothesizing violates the rational prohibition against inconsistent doxastic attitudes. According to the present suggestion, suspending Q entails that one has a doxastic attitude of committed neutrality with respect to Q.  However, insofar as hypothesizing PQ entails treating one possible answer to Q as more plausible than all other potential answers, hypothesizing is a doxastic attitude that is not neutral with respect to Q.  Hence, an agent that simultaneously suspends Q and hypothesizes PQ has conflicting doxastic attitudes, or at least so the objection goes.

However, the preceding line of argument trades on an equivocation.  There are two distinct ways in which the expression “doxastic attitude” has been used in the literature.  Sometimes, it used narrowly to refer to the attitudes of believing, disbelieving, and suspending.  Call this usage of the expression “doxastic attitude” DA-Narrow. However, at other times, the expression “doxastic attitude” is employed to refer more broadly to all attitudes that display a mind-to-world direction of fit, including assuming, supposing, imagining, and hypothesizing.  Call this broader usage DA-Wide When Palmira says that hypothesizing is a doxastic attitude, he clearly has DA-Wide in mind.  However, the prohibition against inconsistent doxastic attitudes only applies to DA-Narrow.  For example, it is common for a philosopher who believes P to assume (for the sake of reductio argumentation) that not-P.  Although the philosopher’s belief and assumption are inconsistent, we certainly would not regard her as irrational for assuming something inconsistent with what she believes.  Why not?  Because assuming P does not entail that one is rationally committed to the truth of P.  Hence, one may assume not-P without violating one’s rational commitment to the truth of P.  The upshot is that the prohibition against inconsistent doxastic attitudes does not apply to DA-Wide.

The salient question, then, is not whether an agent’s doxastic attitudes, understood in terms of DA-Wide, line up but rather whether an agent’s rational commitments line up. We must therefore ask if hypothesizing PQ entails that one is rationally committed to the truth of PQ.  The answer is clearly no.  After all, if hypothesizing PQ entailed being committed to PQ, then hypothesizing PQ at t would entail that Q is settled for one at t and that one was therefore not engaged in inquiry into Q at t.  In sum, holding that hypothesizing entails a rational commitment to the truth of PQ would undermine its status as a penultimate stage of inquiry. 

At best, hypothesizing PQ entails being committed to PQ being more likely to be true than not.  However, I may hold that a proposition is more likely than not to be true and yet suspend about whether it is true.  For example, Suppose the national polls all predict that Trump has a 70% chance of losing the upcoming United States presidential election.  I may consistently believe that it is more likely than not that Trump will lose (based on my available evidence) and yet suspend judgement about whether he would lose.  Consider: during the 2016 presidential campaign, most national polls placed Trump’s chances of losing to Hillary Clinton between 70% and 99%, and projected that she would be the clear winner in states like Pennsylvania and Wisconsin, both of which ultimately went to Trump.  As the eventual 2016 election results illustrate, sometimes the less likely outcome, given the available evidence, is the one that is ultimately realized.  Recognizing this fact, there is nothing inconsistent about believing some outcome to be more likely than not (based on one’s available evidence) and yet deciding to remain agnostic about whether it will obtain.[5]  The upshot is that even if hypothesizing PQ entails a rational commitment to PQ being more likely to be true than not, it does not follow that it is rationally impermissible to suspend Q at t and hypothesize PQ at t.

Moreover, Friedman’s rationale for holding that suspending Q and believing PQ are rationally incompatible is that the former entails treating Q as open while the latter entails treating Q as settled or closed.  However, insofar as hypothesizing is an inquiring attitude, it does not entail having a settled position on some question.  This means that Friedman’s rationale for holding that suspending and believing are incompatible is missing in the case of hypothesizing.  The upshot is that there is nothing on Friedman’s account of the relationship between suspension and inquiry that prevents us from saying that an inquiring agent suspends Q1 at some time t and also hypothesizes P1Q1 at t.  Nor does Palmira supply any other compelling rationale for thinking that suspending and hypothesizing are incompatible.  This means that the defender of the inquirer-as-suspender thesis remains entitled to say that the inquiring agent suspends Q1 at both stages 1 and 2, and that at stage 2 she also hypothesizes P1Q1 alongside suspending Q1.  Indeed, the defender of the inquirer-as-suspender thesis may insist that it is the fact that the inquiring agent not only suspends Q1, but also hypothesizes P1Q1, that distinguishes stage 2 from stage 1.  The upshot is that Palmira’s charge that the defender of the inquirer-as-suspender thesis is unable to register the difference between stages 1 and 2 in terms of the doxastic attitudes each stage implicates ultimately flounders.

This brings us to Palmira’s second line of criticism of the inquirer-as-suspender thesis; namely, that it gives rise to the Unity of Inquiry Problem. Once it is granted that an agent may simultaneously suspend Q and hypothesize PQ, there is no need for the defender of the inquirer-as-suspender thesis to mark the shift from stage 1 to stage 2 in terms of the shift from suspending Q1 and suspending Q2.  Instead, the defender of the inquirer-as-suspender thesis may claim that at both stages of inquiry, the agent suspends Q1 and that the difference between stages 1 and 2 is to be explained in terms of the fact that hypothesizing P1Q1 is absent in the former stage and present in the latter. 

Admittedly, the defender of the inquirer-as-suspender thesis is committed to saying that at whatever point I adopt an inquiring frame of mind towards Q2, I am also suspended about Q2.  However, it is uncontroversial that an agent may suspend two different questions at the same time.  For example, I may simultaneously suspend judgement about where life on earth came from and about whether life on earth originated from non-life.  Hence, the defender of the inquirer-as-suspender thesis may hold that during stage 1, the inquiring agent only suspends Q1, and that during stage 2, they come to suspend Q2 in addition to suspending Q1.  On the present suggestion, what unifies stages 1 and 2 as stages of a single instance of inquiry is that both primarily involve suspending Q1.  However, during stage 2, the agent suspends a secondary question, Q2, in the service of answering the primary question, Q1.  In sum, the agent comes to suspend Q2 in order to facilitate her continued inquiry into Q1, about which she is also suspended. 

The upshot of the preceding analysis is a hybrid view of the second stage of inquiry that includes both hypothesizing P1Q1 and suspending Q2.  According to the view currently on offer, what unifies both stages 1 and 2 as stages of inquiry into a single question is the fact that both stages implicate suspending Q1.  However, once I settle on P1Q1 as the most likely answer to Q1, my inquiring enters a new phase. This new phase is marked by the adoption of the attitude of hypothesizing P1Q1.  Once this hypothesis has been adopted, this initiates a sub-inquiry into Q2 as part of the overarching inquiry into Q1.  This sub-inquiry implicates its own instance of suspension—i.e., suspending Q2. 

However, both hypothesizing P1Q1 and suspending Q2 are attitudes I adopt alongside my suspension of Q1 and are adopted in the service of the inquiry facilitated by my suspending Q1.  In sum, the defender of the inquirer-as-suspender thesis is both able to mark the shift from stages 1 to 2 of inquiry—i.e., in terms of the shift from merely suspending Q1 to hypothesizing P1Q1 and suspending Q2, in addition to suspending Q1—while also avoiding the Unity of Inquiry Problem—i.e., by holding that both stages of my inquiry into Q1 involve suspending Q1.

[1] See and cf. Friedman 2017: 302.
[2] See and cf. Friedman 2019: 303.
[3] Archer (forthcoming)
[4] See Palmira (2018: 5).
[5] This comports with my earlier observation that suspending judgement is the appropriate response to one’s recognition that the available evidence is inconclusive.  It is not merely the appropriate response to the recognition that the available evidence equally supports P and not-P.


Archer, Avery (forthcoming). Agnosticism, Inquiry, and Unanswerable Questions. Disputatio.

Archer, Avery. (2018). Wondering About What You Know. Analysis 78(4): 596-604.

Friedman, Jane. (2017). Why Suspend Judging? Nous, 51(2): 302-326.

Friedman, Jane. (2013a). Question-Directed Attitudes. Philosophical Perspectives 27(1): 145-174.

Friedman, Jane. (2013b). Suspended Judgment. Philosophical Studies 162(2):165-181.

Palmira, Michele. (2018). Inquiry and the Doxastic Attitudes. Synthese.