Grasping a Proposition and Cancellation

Abstract

Recently, Indrek Reiland proposed a new version of the act-type theory of propositions (ATT) in which predication is still committal. However, the Frege-Geach problem can be addressed without resorting to Peter Hanks's cancellation manoeuvre. In this article, I argue that if we take predication as a committal act, we will then have to tackle another problem: non-committal representational acts. I argue that Reiland still needs a notion of cancellation to deal with the latter problem. On this account, he cannot avoid the major flaw he attributes to Hanks's version.

Résumé

Récemment, Indrek Reiland a proposé une nouvelle version de la théorie des propositions comme type d'actes (ATT) dans laquelle la prédication demeure un acte d'engagement. Cependant, le problème Frege-Geach peut être abordé sans recourir à la manœuvre d'annulation de Peter Hanks. Dans cet article, je soutiens que si nous considérons la prédication comme un acte d'engagement, nous devrons alors nous attaquer à un autre problème : celui des actes représentationnels qui n'ont pas de dimension d'engagement. Je soutiens que Reiland a encore besoin d'une notion d'annulation pour traiter ce dernier problème. De ce fait, il ne peut éviter le défaut majeur qu'il attribue à la version de Hanks.

1. Introduction

According to traditional theories of propositions, representation begins with propositions.¹ Let me give an example. Suppose S judges that o is F. S represents o as F via her judgement. This means that S's judgement bears the property of representing o as F. In traditional theories, representationality of S's judgement is explanatorily dependent on the representationality of the object of her judgement, i.e., the proposition that o is F. S's judgement represents o as F because the proposition that o is F represents o as F. In this sense, we say that the proposition that o is F is the primary bearer of representationality, whereas S's judgement is the secondary bearer. In other words, traditional theories of propositions draw on an agent-independent structure ² to explain the representationality of propositions.

Recently, some philosophers have criticized traditional theories with respect to their agent-independent structure.³ They believe that it is mysterious how propositions, traditionally conceived, can represent things as being certain ways. So, traditional theories of propositions cannot provide us with a satisfactory explanation of the representationality of propositions. Therefore, we need a theory of the nature of propositions in which representation no longer begins with propositions. The act-type theory of propositions (ATT) is a new theory about the nature of propositions in which representation begins with agents and their representational acts. In other words, (ATT) draws on an agent-dependent structure ⁴ to explain the representationality of propositions.

Peter Hanks (2007, 2011, 2015) and Scott Soames (2010, 2015) are two leading defenders of (ATT). In their view, the proposition that o is F is an act-type in tokens of which an agent predicates F-ness of o. When S predicates F-ness of o, she represents o as F. So, every token of the proposition that o is F represents o as F. In this way, the proposition that o is F, as the type of those representational tokens, represents o as F. As we can see, in the act-type theory, agents and token-acts of predication are the primary bearers of representationality, whilst propositions are the secondary bearers.⁵

At this stage, the proponent of (ATT) faces an important question: how can S represent o as F by predicating F-ness of o? Soames and Hanks give different answers to this question. As a result, there are two main versions of (ATT) that offer different solutions to the problem of the representationality of propositions.

Indrek Reiland (2019) argues that there are powerful reasons for each of the two leading versions of (ATT). However, each version suffers from a stubborn problem, which is why Reiland suggests a third version that retains those powerful reasons but avoids the problems of the first two versions. My main goal in this article is to show that Reiland's version of (ATT) cannot avoid the most important problem he attributes to Hanks's version.

I will proceed as follows. In Section 2, I will provide an overview of Soames's and Hanks's versions as well as the advantages and flaws Reiland attributes to them. In Section 3, I will give an outline of Reiland's version, showing how this version can preserve the advantages of its rival versions and avoid their flaws. In Section 4, I focus on the major problem Reiland attributes to Hanks's version, i.e., the cancellation manoeuvre. I will argue that Reiland's version cannot avoid the cancellation manoeuvre. I will conclude the article in Section 5.

2. On the Advantages and Flaws of the Two Leading Versions of (ATT)

According to Soames, when S predicates F-ness of o, she entertains the proposition that o is F:

To entertain the proposition that o is red is to predicate redness of o. (Soames, 2014, p. 96)

In the traditional conception of the term, “entertainment” is some special kind of thinking about propositions. For example, within the Fregean theory of propositions, when S entertains the proposition that o is F, she identifies, understands, calls into the mind, or singles out that proposition. S does any of these acts by thinking about the proposition. Soames rejects the traditional conception of entertainment. In his view, entertainment is not some special thinking about propositions; rather, it is to perform/token those propositions:

To entertain a proposition is not, as Frege or the early Russell would have you believe, to think of it in a special way; it is to perform it. (Soames, 2015, p. 18)

The entertainment view of predication is mainly characterized by the rejection of the idea that predication is a committal act. Fregean entertainment is a neutral act. When S entertains, in the Fregean sense of entertainment, the proposition that o is F, she does not take any stance on that proposition. She only identifies the proposition from amongst other propositions in the Platonic world. Similarly, Soamesian entertainment is a neutral act. When S entertains, in the Soamesian sense of entertainment, the proposition that o is F, she does not take any stance on the F-ness of o. She only tokens that proposition.

On Hanks's account, when S predicates F-ness of o, she sorts o with other objects that are F. She does this by applying or attributing F-ness to o. When S mentally predicates F-ness of o, she applies/attributes F-ness to o with a judgemental force. And when she vocally predicates F-ness of o, she applies/attributes F-ness to o with an assertive force:

To predicate the property of being green of an object x is to sort x with other green things. Such an act can be performed, mentally, by judging that x is green, or in speech, by asserting that x is green, or in other ways, e.g., by picking x up and putting it with other things. (Hanks, 2013, p. 560)

Let us use the term “judgement” in an extended sense to cover both mental and vocal predication. In this extended sense, we conclude that Hanks's predication is a judgemental act. The judgemental view of predication is mainly characterized by the idea of predication as a committal act. When S predicates F-ness of o, she judges that o is F. Judgement is a committal act. That is, when S judges that o is F, she is committed to F-ness of o. S is thus committed to F-ness of o by predicating F-ness of o.

At this stage, we can talk about the most important strengths and weaknesses of each of the two versions of (ATT). I start with one version of the Frege-Geach problem. Let us call this version “the first version of the Frege-Geach problem.”⁶ We have a strong intuition that when S judges that if o is F, then o' is G, she neither judges that o is F nor that o' is G. Suppose a theory holds that when S judges a conditional, she judges both the antecedent and the consequent. In that case, the defender of such a theory is faced with the problem of how to explain the above intuition. This is the first version of the Frege-Geach problem.

Hanks's version suffers from the first version of the Frege-Geach problem. According to (ATT), when S judges that if o is F, then o' is G, she must perform the antecedent and the consequent. S performs the antecedent by predicating F-ness of o. Similarly, she performs the consequent by predicating G-ness of o'. In the Hanksian version of (ATT), predication is judgemental. So, when S predicates F-ness of o, she judges that o is F. The same scenario is repeated for the consequent of the above conditional. This means that S performs the antecedent and the consequent by judging them. Thus, when S judges that if o is F, then o' is G, she must judge the antecedent and the consequent. Hanks is accordingly faced with the first version of the Frege-Geach problem.

Hanks replies to the first version of the Frege-Geach problem with his cancellation manoeuvre. He agrees that it is impossible to perform a proposition without judging it. But, when S judges that if o is F, then o' is G, because of the presence of the operator “if,” the judgemental force of the antecedent and that of the consequent are cancelled. We can thus explain the first version of the Frege-Geach problem.

Reiland thinks that Hanks's cancellation manoeuvre is unsuccessful, and so, he cannot explain the first version of the Frege-Geach problem. There are two ways in which cancelled predication might be articulated. According to the first articulation, cancelled predication involves no predication. Reiland (2012) argues⁷ that this articulation faces a dilemma: either the representationality of propositions is also cancelled, or Hanks must endorse Soames's version of (ATT).⁸ Because neither horn is acceptable, Hanks's first articulation of cancelled predication fails. According to the second articulation, cancelled predication is more than predication. Reiland (2019, pp. 147–148) argues that this articulation does not work either. The gist of his argument is that it is unclear what it means for cancelled predication to be more than predication, and any attempt to clarify its meaning fails.

In taking predication as a neutral act of entertainment, Soames was mainly motivated by coping with the first version of the Frege-Geach problem. Let us suppose that S judges that if o is F, then o' is G. S judges this conditional by performing the antecedent and the consequent. S performs the antecedent by predicating F-ness of o. She performs the consequent by predicating G-ness of o'. In Soamesian version of (ATT), when S predicates F-ness of o, she entertains the proposition that o is F. Similarly, when S predicates G-ness of o', she entertains the proposition that o' is G. So, when S judges that if o is F, then o' is G, she entertains the antecedent and entertains the consequent. So, S can judge that if o is F, then o' is G, without judging the antecedent or the consequent. In this way, Soames can avoid the first version of the Frege-Geach problem.

Reiland suggests that the most powerful reason for Soames's version is its avoidance of the first version of the Frege-Geach problem. Moreover, the main problem with Hanks's version is that it cannot satisfactorily address the first version of the Frege-Geach problem.

But this is not the whole story. Hanks's version is also supported by strong evidence, and Soames's version suffers from an important problem. Hanks thinks that Soames's conception of predication (i.e., neutral predication) is incoherent and predication must be committal. Let us suppose that S predicates F-ness of o, where o is not F. From these two suppositions, Hanks (2015, p. 37) argues that:

1. (1) S's token-act of predication is false.

2. (2) S's token-act of predication is incorrect.

3. (3) S made a mistake.

4. (4) S must have taken a position about whether o is F.

5. (5) S's token-act of predication was not neutral.

To justify each move in this argument, Hanks says:

Each move in this argument is based on a triviality. The step from (1) to (2) is based on the fact that truth conditions are correctness conditions. The step from (2) to (3) comes from the fact if you did something incorrect then you made a mistake. The step from (3) to (4) relies on the fact that you cannot make a mistake about whether [o]⁹ is F without taking a position one way or the other about whether [o] is F. Finally, the move from (4) to (5) is based on the idea that if you take a position about some issue then you have not remained neutral about that issue. (Hanks, 2015, p. 37; [footnote added])

Hanks also uses an analogy to strengthen his argument. According to Hanks, a useful analogy for thinking about predication is acts of sorting things into groups. Suppose we have a pile of pencils of different colours. Suppose we decide to put green pencils in a bundle. This is the analogue of an act of predicating the property of being green of pencils. According to Hanks, it is clear that the act of sorting a pencil into a bundle of green pencils is true (correct) or false (incorrect), depending on whether that pencil has the right colour. Moreover, it is clear that by putting that pencil into the bundle of green pencils we are committing ourselves to the pencil's having a green colour. Hanks says:

To remain neutral about whether [the pencil is green¹⁰] would mean not sorting it into the [bundle of green pencils] — it would mean refraining from sorting [the pencil] with the other [green pencils]. But if you don't sort the [pencil] into a group then what you've done (or not done) with it cannot be assessed as accurate or inaccurate.¹¹ Accuracy requires that you sort the [pencil] into a group with other [pencils], but neutrality requires that you refrain from sorting the [pencil]. Combining these features, as Soames does in his concept of predication, leads to incoherence. (Hanks, 2019, p. 1388; [footnotes added])

If this argument works,¹² then neutral predication is incoherent. In (ATT), predication explains representationality of propositions. Nothing incoherent can explain representationality of propositions. So, neutral predication cannot explain representationality of propositions. As a consequence, predication must be a committal act.

According to Reiland, the most powerful reason for Hanks's version is that predication must be a committal act. Moreover, the major problem with Soames's version is that it cannot provide us with a satisfactory explanation of the representationality of propositions. Reiland has a single justification for both claims. He thinks that Hanks's argument against neutral predication delivers the goods:

I find this argument convincing and won't discuss it further here, but proceed with the assumption that it works. (Reiland, 2019, p. 145)

I accept, for the sake of argument, that Hanks's argument works, and neutral predication cannot explain the representationality of propositions. I also accept that Hanks's cancellation manoeuvre is unsuccessful, and so, he cannot address the first version of the Frege-Geach problem. If we accept these claims, we need a version of (ATT) that satisfies two constraints. First, it is committed to the idea that predication is judgemental, and so committal. Second, it can avoid or handle the first version of the Frege-Geach problem.

3. Reiland's Version of (ATT)

Reiland intends to propose a version of (ATT) that preserves the powerful reasons for Hanks's and Soames's versions whilst avoiding their main shortcomings. To do so, he suggests a version in which predication is still judgemental, but the notion of cancellation is no longer needed to address the first version of the Frege-Geach problem.

Reiland begins by contrasting two kinds of acts to reach such a version: propositional and objectual. The paradigm case of the former is the act of believing.¹³ When S believes that o is F, she performs a propositional act in the sense that her act has a propositional content, rather than an object. The paradigm case of the latter is the act of referring. When S refers to o, she performs an objectual act in the sense that her act has a non-propositional object, rather than a propositional content.

One clear criterion for drawing a distinction between propositional and objectual acts is suggested by Alex Grzankowski (2016). According to Grzankowski, a propositional act is sensitive to the truth of the proposition that is the content of the act. However, an objectual act is not sensitive to the truth of the proposition that is the object of the act. For example, when S fears that o is F, her act is propositional because it is sensitive to the truth of the proposition that o is F. It is sensitive to the truth of that proposition in the sense that if that proposition were true, o would be as S fears o to be. By contrast, when S fears the proposition that o is F, as rare as it is to fear a proposition, her act is objectual because it is not sensitive to the truth of the proposition that o is F. It is not sensitive to the truth of that proposition because it is not the case that if that proposition were true, o would be as S fears o to be.

In the next step, Reiland suggests an objectual act of grasping a proposition. He agrees with Hanks that predication is judgemental. So, just like Hanks, he thinks that when S predicates F-ness of o, she performs the proposition that o is F in the sense that she judges that o is F. Yet, Reiland, unlike Hanks, thinks that the capacity to perform (in the sense of judging) the proposition that o is F either comes together with the capacity to grasp that proposition or “at some later point in the phylogenetic development cognizers develop this capacity” (Reiland, 2019, p. 150).

The act of grasping a proposition is objectual. When S grasps the proposition that o is F, she grabs a mental hold of that proposition as an object. There is thus a clear distinction between performing the proposition that o is F as the act of judging that proposition, and performing the proposition that o is F as the act of grasping that proposition. The former is propositional, but the latter is objectual.

If we can accept that there is an act of grasping propositions, then we can address the first version of the Frege-Geach problem without recourse to the notion of cancellation. Suppose S judges that if o is F, then o' is G. S must perform the antecedent and the consequent, to judge the conditional. According to Hanks, S performs the antecedent and the consequent by judging them, but in this case, he faces the first version of the Frege-Geach problem, and as a consequence, he needs his notion of cancellation to address the problem. According to Reiland, S performs the antecedent and the consequent by grasping them, which means that S performs the antecedent and the consequent without judging them. In this way, Reiland avoids the first version of the Frege-Geach problem.¹⁴

Therefore:

1. (1) Reiland agrees with Hanks that predication is judgemental and hence committal. So, his version of (ATT) preserves the most important strength of Hanks's version and avoids the main flaw in Soames's version.

2. (2) Reiland rejects the idea that we must judge a proposition to perform it. We can grasp a proposition without judging it. In this way, Reiland can address the first version of the Frege-Geach problem without an appeal to the notion of cancellation. So, his version of (ATT) preserves the main advantage of Soames's version and avoids the main flaw in Hanks's version.¹⁵

4. Against Reiland's Version of (ATT)

My argument against Reiland's version of (ATT) runs as follows. Hanks needs a notion of cancellation to address the first version of Frege-Geach problem. He also needs the notion of cancellation to address two more problems: the second version of the Frege-Geach problem and the problem of non-committal representational acts. We saw that Reiland could address the first version of the Frege-Geach problem without an appeal to the notion of cancellation. He can address the second version of the Frege-Geach problem with the help of the act of grasping. However, it is unclear how he can handle the problem of non-committal representational acts without a notion of cancellation.

I begin with the second version of the Frege-Geach problem. Let us suppose that predication is judgemental, in which case token-acts of predication will be judgemental. Propositions are types of token-acts of predication. So, propositions inherit the property of being judgemental from their tokens.¹⁶ This means that propositions, as types of judgemental tokens, will also be judgemental.¹⁷

Now consider the following argument:

FG Argument

(FG1) If o is F, then o' is G.

(FG2) o is F.

(FG3) So, o' is G.

The proposition that o is F in (FG1) is not judgemental because it is the antecedent of a conditional. The proposition that o is F in (FG2) is judgemental. Therefore, the proposition that o is F in (FG1) and the proposition that o is F in (FG2) are not identical, in which case FG Argument will not be valid. However, since FG Argument is valid, the proponent of the idea that predication is judgemental faces the problem of explaining the validity of the above argument. This is the second version of the Frege-Geach problem.

Recall that Hanks had the notion of cancellation available to him. Hanks can handle the problem by saying that in the presence of the operator “if” the judgemental force of the proposition that o is F in (FG1) is cancelled. That being so, the proposition that o is F in (FG1) is no longer judgemental. However, when the judgemental force of the proposition is cancelled, its content is not changed. Therefore, the propositional content that o is F in (FG1) and the propositional content that o is F in (FG2) are identical. Accordingly, FG Argument is valid.

Reiland, like Hanks, is committed to the judgemental view of predication. However, unlike Hanks, he rejects the idea that we need the problematic notion of cancellation. For this reason, Reiland must explain how he can tackle the second version of the Frege-Geach problem. To be more precise, Reiland must explain how his manoeuvre of introducing the objectual act of grasping can be relevant to, and handle, the second version of the Frege-Geach problem.

Reiland can address the second version of the Frege-Geach problem as follows.

The problem arises because we are committed to (A):

1. (A) The antecedent of a conditional is not judgemental.

But (A) is not true! The antecedent of a conditional is still judgemental. Therefore, the proposition that o is F in (FG1) is judgemental. So, the proposition that o is F in (FG1) and the proposition that o is F in (FG2) are identical. It follows that FG Argument is valid. Therefore, by rejecting (A), we can address the second version of the Frege-Geach problem.

However, to provide a complete solution to the second version of the Frege-Geach problem, Reiland must explain why the antecedent of a conditional appears not to be judgemental, whilst it in fact is. In order to do this, he can argue as follows. The data that are in need of explanation are the following:

1. (1) The antecedent of a conditional is judgemental.

2. (2) It appears that the antecedent of a conditional is not judgemental.

Regarding (1), the antecedent of a conditional is judgemental because it is the type of judgemental tokens. Let me explain. In what sense does an act-type theorist say that a proposition is judgemental? The answer is obvious: a proposition is judgemental in the sense that it is the type of judgemental tokens. For example, the proposition that o is F is judgemental in the sense that it is the type in tokens of which an agent predicates F-ness of o. When an agent predicates F-ness of o, she judges that o is F. So, every token of the proposition that o is F is judgemental. The proposition that o is F, as the type of judgemental tokens, inherits the property of being judgemental from its tokens. We can thus conclude that this proposition itself is judgemental. The proposition that o is F, whether it occurs in the antecedent position or not, is the act-type in tokens of which an agent predicates F-ness of o. Therefore, whether it occurs in the antecedent position or not, this proposition is judgemental.

Regarding (2), the appearance that the antecedent of a conditional is not judgemental arises because in none of its tokens does an agent judge the antecedent. The proposition that if o is F, then o' is G is the type in tokens of which an agent predicates the material conditional relation¹⁸ of the proposition that o is F and the proposition that o' is G. There is a crucial difference between the proposition that o is F and the proposition that if o is F, then o' is G. In the former case, our target is an object. That is, in every token of the proposition that o is F, an agent predicates F-ness of the object o. But in the latter case, the target is act-types. That is, in every token of the proposition that if o is F, then o' is G, an agent predicates the material conditional relation of the act-type that o is F and the act-type that o' is G. When the target is an object, the agent singles out the object by referring to it. For example, in every token of the proposition that o is F, an agent singles out o by referring to it. When the target is an act-type, the agent singles out the act-type by performing it. For example, in every token of the proposition that if o is F, then o' is G, an agent singles out the proposition that o is F by performing it. According to Reiland, when an agent performs a proposition, she grasps that proposition. It follows that, in every token of the proposition that if o is F, then o' is G, an agent performs the proposition that o is F by grasping that proposition. When an agent grasps the proposition that o is F, she does not judge that o is F. This means that an agent does not judge that o is F in any of the tokens of the proposition that if o is F, then o' is G.

I think the above answer is convincing. If so, Reiland can address the second version of the Frege-Geach problem as successfully as he can address the first version. But is his grasping manoeuvre sufficient for a satisfactory answer to the problem of non-committal representational acts? I don't think so. In the following, I will explain why.

There are some representational acts that are committal. The act of believing, for example, is both representational and committal. That is, when S believes that o is F, she represents o as F and is committed to the F-ness of o. However, not all representational acts are committal. For example, when S imagines that o is F, she represents o as F but is not committed to the F-ness of o. In the following, I will focus on the act of imagining as an example of non-committal representational acts. But what I will say about the act of imagining extends to all non-committal representational acts.

According to (ATT), the ultimate ground of representationality is predication. This means that every representation must be explained in terms of predication in the end. The act of imagining is representational. So, the act-type theorist must say that S predicates F-ness of o when she imagines that o is F. If predication is judgemental, S is committed to the F-ness of o via her imagination. However, S, via her imagination, is not committed to the F-ness of o. So, the proponent of the view that predication is judgemental must explain how S, via her imagination, is not committed to the F-ness of o. This is the problem of non-committal representational acts.

A case in which Hanks needs his notion of cancellation is where he addresses the problem of non-committal representational acts. Hanks can handle the problem by saying that the judgemental force of predication is cancelled in the presence of a non-committal representational act. Therefore, S, via imagining the F-ness of o, is not committed to the F-ness of o.

Reiland is committed to the view that predication is judgemental. So, he must explain how S can imagine that o is F without being committed to the F-ness of o. As we saw earlier, he rejects Hanks's cancellation manoeuvre due to the problems he attributes to the notion of cancellation. Moreover, he introduces the act of grasping to address the first version of the Frege-Geach problem. Thus, Reiland can respond to the problem of non-committal representational acts by appealing to the act of grasping: when S imagines that o is F, she does not predicate F-ness. Rather, she grasps the proposition that o is F. The act of grasping is not committal. So, S, via her imagination, is not committed to the F-ness of o.

The first problem with this answer is that it presupposes that every act of imagining is objectual. However, we have cases in which the act of imagining is propositional. Consider the following cases, for example:

1. (1) S believes that o is F.

2. (2) S hopes that o is F.

3. (3) S fears that o is F.

4. (4) S imagines that o is F.

In these cases, we use “that-clause” to report the intended actions. When we use “that-clause” to report a representational act, the most natural claim we can make about that action is that it is propositional.

Now, look at the following cases:

1. (1) S believes the proposition that o is F.

2. (2) S hopes the proposition that o is F.

3. (3) S fears the proposition that o is F.

4. (4) S imagines the proposition that o is F.

In these cases, we have not used “that-clause” to report the intended actions. Here, the most natural claim we can make is that these actions are objectual.

Let us go back to the answer I suggested Reiland could give to the problem of non-committal representational acts. According to this answer, when S imagines that o is F, she grasps the proposition that o is F. The act of grasping is objectual and non-committal. It follows that the act of imagining is also objectual and non-committal. However, there is a crucial difference between these two cases:

1. (1) S imagines the proposition that o is F.

2. (2) S imagines that o is F.

The natural view is that S's act of imagining in the first case is an objectual one, and her act of imagining in the second case is propositional. The above answer to the problem of non-committal representational acts presupposes that S's act in the second case is also objectual. If Reiland wants to give such an answer to the problem of non-committal representational acts, he must explain why he thinks that all cases of imagination are objectual.¹⁹

But even if we accept that Reiland can explain why all cases of imagination are objectual, there is still a more serious problem. Within Reiland's framework when S imagines that o is F, she does not predicate F-ness of o; rather, she grasps the proposition that o is F. The act of imagining is representational. So, Reiland must explain how S can represent o as F without predicating the F-ness of o, and just by grasping the proposition that o is F. I propose three possible answers and argue that none of them are satisfactory.

According to (ATT), the act of predication is ultimately responsible for representation. Let me call this claim “The Principle of (ATT).” The first answer that Reiland can give to the above problem is to deny The Principle of (ATT). He can answer like this: The Principle of (ATT) is false. We have two primitive representational acts in (ATT): the committal act of predication and the neutral act of grasping. When S imagines that o is F, she grasps the proposition that o is F. The act of grasping is representational. So, S, via her act of grasping, can represent o as F. But this answer has two problems. First, it is ad hoc because it rejects one of the most basic principles of (ATT) to overcome one specific problem of one version of (ATT). Second, it is incomplete at best. We saw how Reiland accepts from Hanks that the neutral act of predication is incoherent and cannot represent anything. So, it is natural to think that, with similar arguments, we can conclude that the neutral act of grasping is also incoherent and cannot represent anything. Therefore, Reiland must explain why those arguments cannot also be applied here.

The second possible answer that Reiland could give to the above problem is to deny that the act of imagining is representational at all. This answer extends to all other representational acts that appear non-committal. Accordingly, the second answer assumes that there are no non-committal representational acts. This answer is question-begging and thus a non-starter.

Lastly, Reiland could answer the above problem in the following way. Let us suppose that S imagines that o is F. S's act of imagination is not representational. It follows that when S imagines that o is F, she does not predicate F-ness of o. However, the object of S's imagination, i.e., the proposition that o is F, is representational. It has the property of representing o as F. When S imagines that o is F, the property of representing o as F is transferred from the object of S's imagination to S's imagination. It explains why S can represent o as F via her imagination.²⁰

At this stage, the question is how the representationality of the object of S's imagination is transferred to S's imagination. Reiland would be best served by relying on the act of performing to answer this question. When S imagines that o is F, she performs the proposition that o is F. When she performs the proposition that o is F, the representationality of the proposition is transferred to S's act of imagination.

Within Hanks's framework, this answer is promising. According to Hanks, S performs the proposition that o is F by predicating F-ness of o. Predication is representational. So, representationality is transferred via the representational act of prediction. However, within Reiland's framework, this answer is unsatisfactory. According to Reiland, S performs the proposition that o is F by grasping that proposition. The act of grasping is not representational. Therefore, it is unclear how the representationality of the proposition that o is F is transferred to S's imagination via the non-representational act of grasping the proposition.

To sum up, Reiland's grasping manoeuvre to address the problem of non-committal representational acts has two defects. First, it must explain why all cases of imagination are objectual. Second, it is unclear how this manoeuvre can explain the representationality of the act of imagination. Reiland still needs the act of predication to explain the representationality of the act of imagination. Therefore, he still needs a concept of cancellation to deal with the problem of non-committal representational acts.

5. Conclusion

If we take predication as a committal act, we must address the first version of the Frege-Geach problem, the second version of the Frege-Geach problem, and the problem of non-committal representational acts. Since Reiland takes predication as a committal act, he must deal with those three problems. He proposes the act of grasping as a way of handling the first version of the Frege-Geach problem. With the same artifice, he can also address the second version of the Frege-Geach problem. However, Reiland cannot address the problem of non-committal representational acts by recourse to the act of grasping. If this is so due to his commitment to a judgemental view of predication, he still needs a notion of cancellation or something to the same effect to address the latter problem. Therefore, Reiland cannot avoid the major flaw he attributes to Hanks's version.