Skip to main content
Article

Thinking Together: Advising as Collaborative Deliberation

Author
  • Joshua Habgood-Coote

Abstract

We spend a good deal of time thinking about how and when to advise others, and how to respond to others advising us. But philosophical discussions of the nature and norms of advising have been scattered and somewhat disconnected. The most focused discussion has come from philosophers of language interested in whether advising is an assertive or directive kind of speech act. This paper argues that the ordinary category of advising is much more heterogenous than has been appreciated: it is possible to advise by asserting relevant facts, by issuing directives, by asking questions, and by other kinds of adviceless advising. The heterogeneity of advising makes speech act-theoretic accounts of advising look like accounts of special cases, which motivates us to look elsewhere for an account of what advising is. I suggest that we think about advising as a distinctive species of joint practical thinking, which I shall call collaborative deliberation, and show how our need for a concept to pick out this kind of joint activity emerges from our practical needs as deliberators. This view helps to illuminate a number of puzzling features of advising, and offers some guidance to be better at advising and at being advised.

Keywords: Advising, speech act theory, philosophy of language

How to Cite:

Habgood-Coote, J., (2024) “Thinking Together: Advising as Collaborative Deliberation”, Philosophers' Imprint 24: 12, 1-19. doi: https://doi.org/10.3998/phimp.3047

863 Views

64 Downloads

Published on
2024-08-26

Peer Reviewed

Introduction

We spend a good deal of time thinking about advising. We seek out advice about relationship problems, household maintenance, and how to write well. We worry about how to give advice to our friends, and how to manage the tension when our parents’ advice slides into telling us what to do. There are professionalised roles for scientific, legal, and financial advisors. Government advisors shape policy and are the first people held responsible when something goes wrong. During the COVID-19 pandemic, we radically altered our lives in response not only to legal restrictions but also to government and scientific advice.

Despite its practical and ethical significance, philosophical discussions of the nature of advising have been scattered and somewhat disconnected.1 This is partly because advising occupies a curious position in the family of speech acts, with connections to both directive and assertive speech acts. Like commanding, advising often involves using the imperative mood to propose a course of action to the hearer. But like asserting, advising can also involve using the declarative mood to make claims about questions relevant to a hearer’s decision.

There are two approaches to explaining the position of advising between asserting and commanding.

The directive approach takes advising to be a directive speech act whose authority is based in knowledge of what is good for the hearer (Hobbes 1998, 2012) (Vendler 1972, p. 41), (Austin 1975, pp. 40–42, 141–2, 155), (Stewart 1978), (Searle and Vanderveken 1985, pp. 202–3), (Bach and Harnish 1979, pp. 40–7), (Hamblin 1987, pp. 10–23), (Wiland 2000a, 2000b, 2021). The best evidence for the directive approach comes from advising involving bare imperatives:

  • 1) Water your seedlings after you plant them!

By contrast, the assertive approach takes advising to be an assertive speech act involving a proposition relevant to the hearer’s decision (Searle 1969, p. 67), (Hinchman 2005), (Sliwa 2012). Although this picture takes advising to be a kind of assertion, it may still retain a connection to action, either via an invitation to take the act of advising itself as a reason for action (Hinchman 2005) or via the assertion functioning as an indirect directive (Nowell-Smith 1954, pp. 146–7). The best evidence for the assertive approach comes from advising involving various declarative formulations:

  • 2) You ought to water your seedlings.

  • 3) Watering your seedlings helps them to root.

  • 4) If I were you, I’d water your seedlings.

The directive and assertive approaches share a commitment to a deference model of advising. According to the deference model, advising centrally involves an epistemic asymmetry, with the basic dynamic of advising being an ignorant advisee seeking a wise advisor in order to defer to the latter’s judgment (see Locher 2006, pp. 5–6). The difference between the two approaches would then be whether advising involves the practical deference we find in orders, or the theoretical deference we find in assertion.

This paper’s goal is to argue that neither the directive nor the assertive approach does justice to the ordinary category of advising. I propose that we think of advising not as a kind of speech act, but as a kind of joint practical thinking, which we pursue by means of different speech acts. In advising, the advisee invites the advisor into his practical problem, and the advisor treats that problem as if it were a shared concern. The advisor does not change the advisee’s question and allows him to make up his own mind. I call this kind of joint practical thinking collaborative deliberation. Joint thinking involves a basic symmetry of position, allowing us to make sense of advising that doesn’t involve epistemic deference (i.e. advising between epistemic peers, and advising by less knowledgeable advisors).

We begin with a hypothesis about the function of our talk and thought about advising: that it answers our needs as deliberators to pool our resources (1.1). This hypothesis motivates the proposal that advising is a distinctive kind of joint practical thinking (1.2). We then consider the heterogeneity of advising (2.1) and the possibility of advising without giving advice (2.2), before showing how the hypothesis explains both this hetrogeneity (2.3) and some of the central features of advising (2.4). We then turn to the case against the assertive and directive views, considering how speech act theorists have understood advising (3.1), before arguing that on all available typologies of speech acts, advising spans the categories of assertives, directives, and askings (3.2). This means that both directive and assertive accounts of advising are mistaken, and that advising cannot be unified with the tools of speech act theory (3.3).

A couple of ground-clearing points.

I shall be pernickety about my use of the terms “advise” and “advice.” “Advise” is a verb marking the activity that is the topic of this paper. “Advice” is a noun with two meanings, referring either to the act of advising or to the object advised. The sentence “Katy’s advice is always so thoughtful”2 can mean either that the way she advises is thoughtful, or that what she advises is thoughtful. “Advice” in its act sense sometimes appears as a generic noun referring to the activity of advising. Although this act/object ambiguity is largely harmless, slippage between the two meanings creates the impression that the act of advising just is the giving of advice. This impression obscures the linguistic fact that “advise” can occur without a complement supplying its content (“Alex advised me”),3 and the non-linguistic fact that we can advise without giving any advice (as I shall argue in 2.2). To avoid this slippage, I use “advice” only in its object sense, and use the gerund “advising” to refer to the activity.

The connection between advising and asserting, directing, and asking means that we need a model for high-level speech act kinds. Until section 3, where we shall need to make more fine-grained distinctions, I use the following basic picture of speech acts, drawn from (Roberts 2018). Sentences in the declarative mood express propositions and are standardly used to assert with the aim of changing the hearer’s beliefs and putting propositions in the common ground. Sentences in the imperative mood express tasks directed toward the hearer and are standardly used to perform directives with the aim of changing the hearer’s intentions and putting tasks on their to-do list. Sentences in the interrogative mood express questions and are standardly used to ask with the aim of putting a question on the shared agenda for inquiry. Uttering a sentence in the declarative mood is the standard way to assert, but it is possible to assert indirectly, and the same goes for directing and asking.

We ought to be careful about claiming that advising is not a kind of speech act. The idea is not that advising does not involve speech acts: it obviously does.4 The idea is that the kinds useful for giving typologies of speech acts will not yield an account of advising, because the everyday category of advising cross-cuts the basic categories of communicative acts. There is no reason to think advising is the only activity masquerading as a speech act, but the arguments we shall canvass do not appear to apply to many communicative verbs.

We rely on English linguistic intuitions throughout. I sometimes note differences in other languages, but I won’t try to mount a proper cross-linguistic survey.

1. Advising and Joint Practical Thinking

1.1. What Is the Point of Advising?

It is often helpful to begin philosophical inquiry about the question what is X? by asking what is the function of our concept of X?.5 Functional approaches to philosophical analysis have a long history but have become more popular following Edward Craig’s Knowledge and the State of Nature (Craig 1990). Craig proposes that the function of our concept of knowledge is to allow us to pool information, and uses this functional hypothesis to illuminate various features of knowledge. How might this approach work in the case of advising?

What is the function of our concept of advising? In his discussion of advising in Practical Reasoning, David Gauthier (1963) offers a hypothesis:

Men [sic] give advice, make recommendations, to assist their fellows with their practical problems. A bread-knife is a device to cut bread; advice is a device to bring one person’s judgement to bear on the problems of another; recommendation is a device to transmit one person’s practical experience to another. (Gauthier 1963, p. 77)

This passage is about the point of advising, but we just as well might take it as a picture of the function of the concept of advising. Just as we need to be able to pool information between people—developing concepts to facilitate and regulate our information-pooling practices—we need to be able to pool our deliberative capacities and resources—developing concepts to facilitate and regulate our pooling of those resources. Craig suggests we think about knowledge from the perspective of an inquirer considering a factual question; I suggest that we think about advising from the perspective of a deliberator considering a practical question.

What might a deliberator need? Allan Gibbard makes a helpful suggestion:

When I ask you for advice, we can say, I try to get you to help me with my thinking, to join with me in thinking what to do. (Gibbard 2003, p. 275)

Let us take up the hypothesis that a deliberator seeks someone who can think through his problem with him. If advising is the kind of activity that answers to the needs of deliberators, we might think of advising as a kind of joint practical thinking.

1.2. Collaborative Deliberation

The idea that advising is a kind of joint thinking is suggestive but needs sharpening. Following de Kenessey (2020), let us think about the standard case of joint practical deliberation (Bratman 1992, Gilbert 2009) as a joint activity with three features:

  • i. It aims at producing a joint decision.

  • ii. It aims to answer a question about what the agents should do together.

  • iii. It is responsive to reasons shared between the agents.

Consider a group of runners deliberating about the route and pace for a Sunday run. They aim to reach a joint decision (keep to 7.30 pace, turn around when you reach the bridge), which will constrain their downstream planning both as individuals and as a group. This decision answers a shared question about what everyone should do. And the reasons relevant to this decision will be a combination of individual reasons (which events people are training for) and the shared norms of the group (run at the slowest desired pace).

Advising is not just any kind of joint practical thinking: the runners are not advising one another when they plan their long run. So how might we fix on a suitable kind of shared thought to identify advising with?

A first thought is that advising is joint practical thinking concerning what the advisee should do in light of his reasons. This won’t do. Consider a case where you and I go to dinner every week, alternating who chooses the restaurant. One week I call you to ask for help deciding between Thali and Root. You talk me through our preferences and dietary restrictions and then leave the decision up to me. Have you advised me? Plausibly, yes: we have pooled our deliberative resources, and you have not ordered or threatened me. But the reasons involved as well as the resulting intention, are shared. If you had got bored and said “Let’s go to Root!” as a means to express a decision, you would have gone beyond advising by taking the joint decision into your own hands. This suggests that advising involves a kind of joint thinking that leaves the responsibility for deciding up to the advisee. Gauthier expresses this point nicely:

To advise, or to recommend, is to assist someone in making a decision or choice, in solving a practical problem. The decision, and the problem, belong to [the] advisee. If the speaker seeks to make the decision, and to impose it on [the] advisee, then he is no longer advising, for he is no longer (just) assisting. He is treating the problem as a joint one, to be faced collectively, rather than the advisee’s own problem. (Gauthier 1963 p. 70)

There are two ways to fill out this idea.

The first is to say that advising involves the off-line deployment of the advisor’s deliberative capacities in order to deal with another person’s practical problems, without either the authority to generate preemptive reasons (as in ordering), or the standing to make shared proposals (as in joint deliberation).

The second is to say that advising involves a pretense of joint deliberation (see Portner 2018 p. 310). The advisee and advisor enter a pretense in which they treat his problem as if it were a shared concern, deploying their deliberative capacities together in a sort of make-believe of shared deliberation. This pretense has an asymmetric character: both advisee and advisor are pretending to engage in shared deliberation, but the advisee genuinely engages in practical deliberation about what to do, while the advisor does not. Let us call this kind of joint practical thinking, in which the advisor deploys her deliberative capacities off-line as part of a pretense of shared deliberation, collaborative deliberation.

Collaborative deliberation should be distinguished both from joint practical deliberation and from merely talking about a problem.

Consider a case where advising shifts into joint practical deliberation. 6 Tariq asks Hannah about the best way to train for a marathon. She develops a training and nutrition plan, and he decides to follow it. To Tariq’s surprise, Hannah starts calling him early in the morning to ensure he wakes up for training sessions, dropping off meals at his house, and booking in sessions with the physio. Everything she is doing might be helpful, but by taking on the normative obligations associated with a joint decision, Hannah has shifted from advising into joint planning. Tariq could tell her to buzz off. Unlike proposals in shared deliberation—where both parties form an intention to do something—and orders—where the speaker at least assumes an obligation not to interfere with the hearer—in collaborative deliberation the advisor comes away with no commitments.7

Next, consider the distinction between advising and merely talking about a problem. Zahir gets home from work and starts complaining to Marta about one of his colleagues. As a good friend, Marta chips in with suggestions about how Zahir might manage the relationship better. Rather than welcoming these suggestions, though, Zahir gets annoyed and tells Marta to stop trying to problem-solve. Here’s one way to understand this conversation. Zahir is just venting: he just wants to let off steam about his colleague. Marta understood Zahir to be instigating a shared conversation about how to deal with his colleague, so she engages in the pretense of shared deliberation. In venting, the shared goal is not to resolve the practical problem under discussion; in collaborative deliberation, it is.

A worry. The importance of the locution ‘if I were you…’ to advising might suggest that advising is a pretense of individual deliberation. There is a style of advising that involves thinking through the advisee’s problem as if it were your own, without taking much input from the advisee. However most cases of advising involve a back-and-forth where the advisee’s input is central, and this exchange can occur only in a pretense of joint—rather than individual—deliberation.

Collaborative deliberation involves the pretense of joint deliberation, but unlike standard cases of joint practical thinking, it does not aim at a joint decision, involve a shift to a shared practical question, or a shift to a shared deliberative perspective.

Although collaborative deliberation aims to help the advisee form a decision, it is insulated from the intentions of both advisee and advisor. Within the pretense, joint deliberation aims to solve the advisee’s problem, but neither the advisee nor the advisor commits to that plan outside the pretense, and the deliberative responsibility remains with the advisee. There is a further step the advisee must take in deciding to follow the advice, and—as the Tariq/Hannah example shows—the advisor should not come away with any commitments about what the advisee is to do.

Collaborative deliberation involves pretending that a problem faced by the advisee is a shared concern, but the problem itself remains unchanged. If the advisee is working out how he ought to break up with a partner, the problem is how he should break up with them, not how he and the advisor should break up with them. Many of the features of the problems we seek advice on depend on them being our problems. If the advisee’s problem already concerns what a group of people should do—either because the advisee is a group, or because an advisee has responsibility for a joint decision, like in the dinner case—collaborative deliberation will concern a shared question. But that question remains unchanged by the shift to an advisory context.8

Within the pretense of collaborative deliberation, the relevant beliefs and preferences are the advisee’s. There is space within the pretense for the advisor to try to persuade the advisee to change both his beliefs and preferences (see 2.4.1), but the starting point is the advisee’s point of view. In cases where the advisee makes a decision from a perspective not his own, perhaps because of role responsibilities or responsibility for a shared question, advising will take place from those perspectives.

We can now offer an account of advising:

COLLABORATIVE DELIBERATION: A advises B just in case A and B together engage in joint thinking about a problem P that is B’s responsibility, within the scope of a pretense where P is a shared concern for A and B, while both the nature of that problem and the reasons bearing on it remain unchanged by the pretense, and B retains responsibility for resolving the problem.

This account takes advising to be a kind of joint thinking, and not a kind of speech act. As noted above, advising will usually (perhaps always) involve speech acts: collaborative deliberation is typically a linguistic pretense. It won’t do any harm to think about advising as the kind of speech act involving moves within collaborative deliberation, as long as we remember that this is not the standard way to individuate speech acts.

2. The Character of Advising

In this section, we consider the features of advising. We first survey the diversity of advising reports (2.1), before defending the claim that it is possible to advise without offering advice (2.2). We then show how the hypothesis that advising is collaborative deliberation can explain both the diversity of advising (2.3) and several of its distinctive features (2.4).

2.1. Advising Reports

In English, the first-person performative “I advise …” is reserved for strong recommendations (Diedrich and Holn 2012), so to get a sense for the ordinary extension of “advise,” we are better off starting with third-person reports. There are four basic kinds of advising reports:

  • (4) Laura advised Robin to take up running.

  • (5) Ruth advised Mark that there were frogs on the cycle path.

  • (6) Katy advised Jack where to get a sourdough pizza.

  • (7) Heather advised caution.

In (4) the complement of “advise” is an infinitive. This report would be appropriate if Laura had uttered either a bare imperative—“take up running!”—an ought claim that had imperative force—“you should really take up running”—or a performative sentence involving “advise” together with either an imperative, ought claim, or infinitive—“my advice is: take up running!”; “I’d advise that you ought to take up running”; “my advice is to take up running.”

In (5) the complement is a declarative phrase. This report would be appropriate if Ruth had uttered a simple declarative phrase—“There are frogs on the cycle path.” Although this assertion may function indirectly as a recommendation (Nowell-Smith 1954, pp. 146–7), it need not. If Mark does not know that the frog species on the path is endangered, or if he fails to realize the risk of running over frogs, he might not recover a recommendation to avoid the cycle path.9 Some writers claim that advisory assertives always indirectly recommend—see (Wiland 2021, pp. 117–18)—but there is plenty of advising involving assertions without indirect directives. A financial advisor might advise by asserting facts relevant to investment decisions without making any indication about where to invest.

In (6) the complement is an interrogative phrase. This kind of report would be appropriate either if Katy asserted a proposition answering the question of where to get a sourdough pizza, or if Jack asked where he could get a sourdough pizza and Katy responded with an imperative—“go to Flour and Ash!” Although advising typically concerns practical questions, both information and directions are appropriate responses.

In (7) the complement is an abstract noun. This kind of report would be appropriate if someone went to Heather for advice, having settled on doing something, but without having formed a fine-grained plan for how. Heather’s advice concerns the best way to climb Death Mountain, which might be expressed either in a declarative—“it’s important to be cautious,” or in an imperative, “be cautious!”

It is important that advising can be reported with both infinitival and declarative complements (Vendler 1972, pp. 20–1). Although there are other verbs—notably “tell”—that pattern this way, it is not a common feature of speech act verbs. We might distinguish two kinds of advising: advising-that (advising where the advice can be reported with a declarative complement), and advising-to (advising where the advice can be reported with an infinitival complement) (Searle 1979, p. 28). These categories overlap: a modal statement made with the force of an imperative could be reported in either way, as could a non-modal statement functioning as an indirect directive. But the categories are not identical. The provision of information without an explicit or implicit recommendation is advising-that but not advising-to, and a bare imperative without any justification is advising-to but not advising-that.10

I take the distinction between advising-that and advising-to to be good preliminary evidence that advising includes both assertive and directive speech acts. While this might seem a surprising claim, it has been widely accepted by both philosophers11 and sociolinguists12 writing about advising.

2.2. Advising Without Advice

Besides advising by uttering imperatives and declaratives, it is possible to advise just by asking questions (Gauthier 1963, p. 50). One might think of an episode of advising composed entirely of interrogatives—without any indirect assertions or directions—as an instance of adviceless advising (still bearing in mind our pernickety use of “advice”).

Consider a case:

MOVING AWAY: Fred is considering moving away from his hometown for an unspecified number of years, and his partner is committed to staying. Fred worries whether they should stay together. He goes to his friend Alex for help. She asks him about various aspects of his relationship, whether he has thought about the different ways to handle a long-distance relationship, and what his partner thinks about the move. But at no point does she either offer any recommendations about what to do or assert any relevant propositions.

Does Alex offer Fred any advice? Asking a question can give advice when it indirectly recommends a course of action, or conveys information. Asking a question can also be a preparatory move to establish the relevant options.13 Imagine Alex is simply posing questions as suggestions for issues Fred might consider, without a clear view about what Fred should do. Then I think that we should say that Alex is not giving Fred any advice.

Does Alex advise Fred? Fred has come to Alex for help with a practical question, and Alex has engaged a kind of joint activity which helps Fred to make up his mind. This seems like good evidence that Alex advised Fred. In addition, many sociolinguists are happy to classify question-asking as a subspecies of advising.14 So we have decent first-pass evidence that this case involves advising.

There are two worries about adviceless advising. First, one might worry that Fred has helped Alex without advising him. What is the difference between asking Fred questions, bringing him tea, or boosting his mood with a game of tennis? Although the distinction between advising and helping is murky, Alex seems to be doing more than just helping by engaging in a joint activity. If playing a game of tennis helps Fred deliberate well by boosting his mood, at most Alex deserves credit for enabling good thinking. By contrast, if asking him a string of questions helps Fred deliberate well, then Alex deserves partial credit for the quality of Fred’s decision.

Second, one might worry about the awkwardness of talking about adviceless advising. It does sound odd to say:

  • (8)?Alex advised me, but she didn’t give any advice.

The question is whether this awkwardness is generated by a linguistic habit of specifying the content of advising, or by the impossibility of adviceless advising. Similar sentences for activities close to advising are less problematic. Consider counseling, in its non-therapeutic sense:

  • (9) Alex counseled me, but she didn’t offer any counsel.15

Although some English speakers report qualms about (9), reflecting on counseling reminds us that there are plenty of ways to counsel without offering counsel: going through the options, telling stories, applying decision-making heuristics, and so on. If we allow counsel-less counseling, we should allow adviceless advising.

Adviceless advising is not just possible; it is valuable. It respects autonomy, enables skill learning, and avoids social awkwardness.

In some domains, it is important that we make decisions for ourselves. Although offering advice does not impugn autonomy in the way that commanding does, it does risk sliding into a kind of joint decision-making that would be inappropriate for, say, relationship decisions. Adviceless advising mitigates that risk, focusing on what the advisee should be paying attention to, and not on what he should do.

We care about doing the right thing and being able to deliberate well. Directive advice gives a shortcut to the right action, but does not help us learn in the way that non-directive advising does (Locher 2006, p. 193).16 When an advisor works through a decision with an advisee by laying out the relevant practical issues—perhaps even withholding a portion of their knowledge—it provides an opportunity to practice good deliberation through joint activity.

Adviceless advising can also avoid social awkwardness. Ordinarily we want issues about our personal lives to be at the core of our epistemic territory: those topics we are authoritative and competent about (see Nagel 2019). Soliciting or accepting directive advice or even assertions about personal topics runs the risk of setting up a context in which we presuppose that an advisor knows more than we do about our business (see Heritage and Sefi 1992, p. 410). Adviceless advising can function as a face-saving device, allowing an advisor to convey her message without undermining the advisee’s social-epistemic standing (Locher 2006, chapter 6, chapter 9).

2.3. Explaining the Diversity of Advising

We have seen that advising is a much more heterogenous activity than we might have thought. Our ordinary category of advising spans assertive, directive, and asking-type speech. Neither the assertive or directive approaches are well-placed to explain this diversity, but let us leave their attempts to deal with it until the next section, and consider how the view that advising is collaborative deliberation can account for it.

Above we said that the prototypical deliberator is after someone to help him with a practical problem. What kinds of help might he seek?17 We might distinguish four kinds of problems faced by our prototypical deliberator. First, he might be out of his depth, meaning that responsible deliberation is beyond him. This deliberator needs a bare recommendation and faces the problem of ensuring that the person giving the recommendation is trustworthy and has his interests at heart. Secondly, he might be in an ignorant situation, lacking factual information relevant to his decision, needing someone to provide him information about options, outcomes, and so on. This deliberator is rather like a Craigian inquirer. Thirdly, the deliberator might be in a high-stakes situation, having both information and deliberative skills but needing someone to work through the decision with him. Fourthly, he might be in the position of the novice, who is interested in developing his deliberative skills.

These different kinds of problem call for different kinds of help: deliberators who are out of their depth need simple recommendations (advising-to); ignorant deliberators require information (advising-that); and high-stakes deliberators and novices require someone to think through a decision with them (adviceless advising). By distinguishing different kinds of deliberators, we can predict both the existence and the importance of different kinds of advising. These four deliberators are ideal cases. They may not exhaust the options, and a real-life advisee will often face several problems, requiring a mix of different kinds of advising.

2.4. Explaining Features of Advising

Identifying advising with collaborative deliberation helps to explain four important features of advising: its distinctive modal force, the fact that advisory imperatives do not create reasons, the possibility of advising between equals, and the badness of unsolicited advising.

2.4.1. Advisory Modals

We have seen that advising can involve both imperatives and modal claims. Both kinds of advisory speech can be reported using an infinitival construction, which has an implicit modal operator. An important feature of advising is that sentences can be used to advise only if they articulate a particular kind of modal force. An imperative or ought claim that appeals to a hierarchical social system, or exclusively to an advisor’s desires and goals, is not advising (Portner 2007, p. 356). Characteristically, advising involves either bouletic modality indexed to the advisee’s desires (“given that you love aerobic exercise, take up running!”), teleological modality indexed to the advisee’s goals (“given that you’re trying to get fit, take up running!”), or deontic modality associated with a system of rules the advisee is already committed to (“given that school requires you to take a sport, take up running!”18). If advising is a kind of pretense in which advisor and advisee treat the advisee’s problem as if it were a shared concern, then the reasons relevant to that problem will not be the advisor’s or shared reasons. They will be the advisee’s alone, hence the close relation between advising and bouletic and teleological modality.19

That advising appears indexed to the advisee’s desires and goals does not mean that all advising takes place from within the advisee’s belief and preference sets (see Nowell-Smith 1954, pp. 155–7, Gauthier 1963, pp. 54–6, Andreou 2006). Advisors can attempt to persuade advisees to change their beliefs or preferences, and may issue directives presupposing beliefs or preferences the advisee does not currently have.20 If these attempts are unsuccessful, or the attempt is viewed as futile, then the advisor may end up issuing advice indexed to beliefs or preferences that she does not share.21

2.4.2. Advising and Reasons

Unlike ordering, which creates normative facts through the exercise of authority, imperatival advising—like warning and recommending—is answerable to prior normative facts (although advising can have normative consequences by making these facts available to the advisee). This point goes back at least to Hobbes:

Now COUNSELL is a precept in which the reason of my obeying it, is taken from the thing it self which is advised; but COMMAND is a precept in which the cause of my obedience depends on the will of the Commander. For it is not properly said, Thus I will, and thus I Command, except the will stand for a Reason.” (Hobbes 1998, S XIV 1)22

The fact that imperatival advising rests on prior normative facts means that advising can be subject to epistemic challenges. Consider a European parent uttering sentences (10) and (11) to a teenager going to a party where curfew is under the scope of the household rules, and drinking is not:

  • (10) Be home by 10:30! [ORDER]

  • (11) Don’t drink more than two beers! [ADVICE]

The teenager can challenge (10) only by appealing to his parent to change her mind. By contrast, he can legitimately challenge (11) by asking what would be so bad about drinking more than two beers. If he can argue the point, then his parent would have to retract her imperative in a way she would not have to do with an order. Relatedly, it is possible to order someone to do something when it is common knowledge that it is not the best thing to do (“just do what I say and come home by 10.30!”), but advising must be done under the guise of the good.

This difference between advising and ordering is neatly predicted by the idea that advising is collaborative deliberation. Although advisors may employ the linguistic markers of orders or joint decisions, when they enter into collaborative deliberation, the issue to be resolved remains the advisee’s problem. The advisor goes wrong if she changes this problem by creating new normative reasons.

2.4.3. Advising Without Deference

There are cases in which we go to advisors because we need knowledge: recall the prototypical deliberators who are out of their depth, ignorant, or novices. But in many cases we are advised by people who are just as knowledgeable as us, or who even might know less than us. In MOVING AWAY, Fred need not think Alex is an expert on relationship matters, and she does not need to present herself as such to offer useful contributions. Recall the situation of the high-stakes deliberator from 2.3: sometimes we just want someone to think through a problem with, and it does not matter that they know more or are more deliberatively skilled than we are. The idea that advising is collaborative deliberation nicely predicts this possibility: joint thinking puts the two parties on equal standing.

3. Advising Is Not a Kind of Speech Act

Let’s take stock. We have framed a hypothesis about the function of the concept of advising, and used that hypothesis to motivate a view about what advising is. We argued that this view illuminates a wide range of advising’s features, including its heterogeneity. With the positive case in place, we turn to the treatment of advising by speech act theorists in order to argue for two claims: that the directive and assertive views are mistaken, and that advising is not a kind of speech act. The central part of the negative case against the assertive and directive views is that these accounts are unable to explain the heterogeneity of advising.

We begin by surveying how speech act theorists have classified advising (3.1), before showing that whatever theory of speech acts is correct, advising spans the distinction between types of speech act (3.2), and arguing argue that this fact gives a knock-down argument against the assertive and directive accounts of advising (3.3).

3.1. Speech Act Typologies and Advising

Speech act theorists have classified advising in various different ways.

Austin’s classification of illocutionary acts focuses on the performative uses of speech act verbs. He groups advising with exercitives, which he glosses as “the giving of a decision in favour of or against a certain course of action, or advocacy” (Austin 1975, p. 154). Vendler adds grammatical criteria to Austin’s typology, keeping advising in the category of exercitives on the grounds that advisory performatives involve either an infinitival complement or a declarative ought statement (Vendler 1972, pp. 20–1).

Bach and Harnish group speech acts by which mental states they express. They group advising with advisories, which they classify as directives on the grounds that they communicate an attitude toward a prospective action. They offer a definition of this category (see Hinchman 2005):

As for advisories, what the speaker expresses is not the desire that H do a certain action but the belief that doing it is a good idea, that it is in H’s interest. S expresses also the intention that H take this belief of S’s as a reason to act. The corresponding perlocutionary intentions are that H take S to believe that S actually has the attitudes he is expressing and that H perform the action he is being advised to perform. (Bach and Harnish 1979, p. 49)

In a footnote (1979 fn3), they observe that advising can be performed by either imperatives or declaratives, and contend that in the latter case, advising involves an indirect speech act.

Searle classifies speech acts based on their constitutive rules. An early time-slice of Searle classifies advising as an assertive speech act:

Contrary to what one might suppose advice is not a species of requesting. It is interesting to compare “Advise” with “Urge” “Advocate” and “Recommend.” Advising is not trying to get you to do something in the sense that requesting is. Advising is more like telling you what is best for you. (Searle 1969, p. 67)

He later changes his mind. Searle (1979, p. 28) distinguishes between advising-to and advising-that, and Searle and Vanderveken (1985, pp. 202–3) classify advising as a directive that can take direct imperatival or indirect declarative forms.

The disagreement among speech act theorists about how to classify advising is no surprise if advising involves both assertive and directive speech acts. It is not hard to find evidence for both views. Searle and Bach and Harnish’s appeal to indirect speech acts suggests a strategy for the supporters of the assertive and directive pictures to explain the diversity of advising. The idea would be to unify advising not by primary illocutionary acts, but by a combination of primary and indirect speech acts. For example, the assertive picture will maintain that cases of advising-to are indirectly assertions of an underlying normative fact supporting the course of action. (The assertive view might think about advisory imperatives as involving making as if to direct (see Harris 2014, pp. 106–11, 2021 fn4)). Similarly, they might think of interrogative advising as the indirect assertion of a relevant normative or non-normative fact.

3.2. Speech-Act Theoretic Features of Advising

We now turn to typologies of speech acts to argue that advising cannot be contained within the categories of assertives, directives, or askings. Typologies of speech acts appeal to the following kinds of properties to distinguish the basic categories of speech acts:

  • 1. Illocutionary point (the essential purpose of a type of speech act) (Austin 1975, Searle 1979, pp. 1–29, Searle and Vanderveken 1985);

  • 2. The direction of fit of the content expressed by the speech act (whether word-world, as in the case of assertives, or world-word, as in the case of directives) (Searle 1979, pp. 1–29, Searle and Vanderveken 1985);

  • 3. The kind of content expressed by the sentences uttered in making the speech act (whether propositional content, imperatival content, or interrogative content) (Roberts 2018);

  • 4. The effects of the speech act on the conversational scoreboard (assertives update the common ground, directives update the audience’s to-do list, askings update the questions under discussion) (Stalnaker 1978, Portner 2007, Roberts 1996/2012);

  • 5. The mental states expressed by the speech acts (with assertives expressing belief, directives expressing desire or intention, and interrogatives expressing the desire to know or the intention to discover) (Bach and Harnish 1979).

Let’s go through these features in turn, considering how advising displays the features which these views associate with assertives, directives, and askings.

3.2.1. Illocutionary Point

Illocutionary point corresponds to the essential point of a type of speech (Searle 1969, chapter 3, 1979, pp. 2–3, Searle and Vanderveken 1985, chapter 9). According to Searle (1979, pp. 12–14), if a speaker is asserting, she is committing herself to the truth of the proposition asserted; if she is commanding, she is attempting to get the hearer to do the action commanded; if she is asking, she is attempting to get the hearer to answer her question. For Searle, illocutionary point is characteristic of a high-level speech act type—such as assertives—which is combined with various other features, including degree of strength, preparatory conditions, and mental state expressed. Ths combination produces an illocutionary force that is characteristic of a particular speech act, such as telling (Searle 1979, p. 3).

Advising involves various illocutionary points. Advising-that involves the speaker committing herself to the truth of a proposition. Advising-to involves something close to attempting to get the hearer to do something, although unlike commanding the grounds for this attempt are the advisee’s antecedent reasons. Interrogative advising is also a little unusual. The standard case of asking a question has often been understood as a request that the hearer answer a question, together with an expression of the desire to know the answer (Bach and Harnish 1979, p. 47, Searle and Vanderveken 1985, pp. 199–200). In the case of advising by asking a question, the point is often to help the advisee come to know the answer, or to propose the question as a target for shared consideration.23

The illocutionary point of the direct speech acts involved in advising are diverse. But what about the illocutionary points of the indirect speech acts involved in advising? It is not clear whether we can find evidence for enough indirect speech acts for either assertive or directive version of this strategy to work. If I ask you to advise me about a possible career change, and you ask me “What parts of your job do you like?” and then tell me “The job market is tough everywhere,” neither piece of advising necessarily involves a recommendation. To make the strategy of unification via indirect speech acts work for either view, we would have to either find evidence that we are systematically engaged in one or other indirect speech act while advising, or commit to drastically restrict the category of advising.

3.2.2. Direction of Fit

The direction of fit associated with a sentence corresponds to the way a sentence relates to the world (Anscombe 1957, pp. 56–7, Searle 1979, pp. 3–4). Some sentences represent the world, meaning a mismatch between word and world is associated with a fault in the sentence. Other sentences change the world to bring it in line with words, meaning a mismatch involves a fault in the world. Assertives have word-to-world direction of fit, and directives have world-to-word direction of fit.24 Advising-to involves world-to-word direction of fit: advising someone to do something functions to get them to pursue the course of action recommended, and if the advice is not taken, the fault lies (primarily) with the advisee and not with the advisor or her word. By contrast, advising-that involves word-to-world direction of fit: advising-that functions to represent a relevant part of the world, and if the advice is wrong, the fault is with the advisor’s word, not the world.

3.2.3. Kind of Content

Although there is a tradition of assimilating all meaning to propositional content (Belnap 1990), an emerging paradigm within philosophy of language, and linguistics—synthesized in (Roberts 2018)25—claims that different grammatical moods track different kinds of content. If we take declarative sentences to express propositions, imperative sentences to express directed tasks (Portner 2007), and interrogative sentences to express questions (Groenendijk and Stokhof 1984), the fact that advising involves declaratives, imperatives, and interrogatives means it involves three different kinds of content.

3.2.4. Scoreboard Effects

We might think about different kinds of speech acts in terms of the ways they update conversational context, understood as a complex of shared mental states. Roberts (2018) develops this view into a typology of speech acts. She suggests that the basic categories are assertions, which are proposals to adopt shared beliefs, and suggestions, which are proposals to adopt shared intentions. Suggestions break down into directions, which are proposals to adopt shared intentions to do things, and interrogations, which are proposals to adopt shared inquisitive plans. On this view, advising would split between the major categories: advising-that is a proposal to update the common ground with information relevant to the advisee’s decision; advising-to is a proposal to update the advisee’s to-do list with a particular action in response to his practical situation; and interrogative advising is a proposal for the hearer to answer a question.

3.2.5. Mental States Expressed

Following Bach and Harnish, we might distinguish speech acts via the mental states they express. The standard test for the relation between speech acts and mental states is the Moorean sentence. If combining a performative speech act with a denial of a mental state is strange in the same way as “it’s raining, but I don’t believe it is” is—anomalous, but not inconsistent—then we have evidence that the speech act expresses the mental state.26 I take it that assertive speech acts express belief, and directive speech acts express either desires or the belief that the direction will be successfully followed.

Focusing on advising-that and advising-to and the omissive version of Moore sentences gives us six sentences to consider:

  • (12) #I advise you that the train is leaving in ten minutes, but I don’t believe that it will.

  • (13) I advise you that the train is leaving in ten minutes, but I don’t want you to catch it.

  • (14) I advise you that the train is leaving in ten minutes, but I don’t believe you are going to catch it.

  • (15) #I advise you to go to Dotori, but I don’t believe that you ought to go.

  • (16) I advise you to go to Dotori, but I don’t want you to go.

  • (17)?I advise you to go to Dotori, but I don’t believe that you will go.

(12) is weird in the same way as the original Moore sentence for belief. This is unsurprising, given the suggestion that advising-that is a kind of assertion. Neither (13) nor (14) is strange. If the first conjunct is a simple assertion (“the train is leaving”) without any associated indirect speech act, it would be reasonable to combine it with the statement of lack of desire, or disbelief in another proposition.27 (15) appears odd in the Moorean way. It is possible to advise someone to do something you do not believe they ought to do, but uttering this sentence is very strange. (15) combines a recommendation with a denial of belief in the normative grounding of that recommendation. If, as suggested in section 2.3, advising-to is a directive grounded in knowledge of the advisee’s good, we can explain the oddness of (15) by taking it to combine a speech act with the denial that one is properly positioned to undertake that act.

I suggest that sentences like (15) will be odd for all directive speech acts (see Portner 2007), although the explanation will vary. Consider orders:

  • (18) #I order you to take a seat, but I don’t believe that you ought to.

(18) is odd because the expected upshot of the order in the first clause is that the task of taking a seat is added to the hearer’s to-do list, making it the case that he ought to take a seat. Thus, although the reason for the oddness of (15) depends on certain distinctive features of advising, the general pattern occurs for all directives.28

(16) is fine. Although we normally expect advisors to desire our good, an advisor can recommend a course of action promoting the advisee’s goals while frustrating the advisor’s. Advising is not unusual in this respect: warnings, exhortations, and suggestions are all indexed to the hearer’s goals and desires, opening the possibility for divergence between a direction and speaker’s desires.

(17) seems odd, especially if we replace the first clause with an imperative:

  • (17*)?Go to Dotori, but I don’t believe that you are going to go.

(17*) is related to what Mandlekern (2021) calls practical Moore sentences: sentences combining an order with an indicative that leaves open whether the order will be followed.

  • (19) #You must close the door, but I don’t know whether you will.

Although it is tempting to assimilate (17) and (17*) into this category, we can establish contexts where they are acceptable. Consider a case where a foodie utters (17) to give unsolicited advice to her disorganized friend (imagine the second clause spoken in an exasperated tone). In this case (17) is perfectly comprehensible, if a little rude. This is not a surprise: practical Moore sentences involve orders, and we have seen various ways that advising and ordering differ. Underlining this difference, note that combining an order with a declarative that leaves open whether it will be followed is odd (Mandlekern 2021), but analogous sentences for advising are fine:

  • (20) I advise you to go to Dotori, but you might not go.

  • (21) Go to Dotori! But it’s up to you to decide where to go.

The residual weirdness around (17) and (17*) owes to the fact that it is usually not helpful to advise someone to do something that they are not in a position to do. These sentences give bad advice, but do not involve pragmatic contradictions.

The data in (12)–(17) is complicated, but the underlying pattern is that advising displays Moorean sentences that are distinctive of both assertive (13) and directive (16) speech acts. It appears that advising can express both belief in the proposition asserted, and a belief about what the hearer ought to do.

3.3. The Place of Advising in Speech Act Theory

We have seen that advising displays the characteristic features of assertive- and directive-type speech acts. There are four possible reactions to this evidence:

  • i. Advising is not a unified category; the word “advise” is ambiguous.

  • ii. Advising is a unified speech act, but we frequently make false claims about advising.

  • iii. Advising is a unified speech act, but it is systematically pursued in an indirect way.

  • iv. Advising is not a unified speech act; it is unified by some other feature.

The view that “advise” is ambiguous between advising-to and advising-that seems to be Searle’s 1979 position. If “advise” were ambiguous, we would expect to detect it using the standard tests for ambiguity (Zwicky and Sadock 1975), but applying the tests does not detect ambiguity.29

An alternative way to maintain the assertive or directive view would be to adopt an error theory for troublesome advising reports. Someone committed to the directive view might think that sentences like I advise you that…” and “Alex advised me that…” are inappropriate uses of force markers, in the sense that what speakers are doing is not advising but merely helpfully asserting. There are two problems with both assertive and directive error theories. First, it is unattractive to ascribe widespread error to sociolinguists, who classify both assertives and directives under advising (see footnote 14). Second, the proponent of this view would need to find a good reason why we ought to accept an error theory about advising-that, rather than an error theory about advising-to (or vice versa).

The view that advising is unified through indirect speech acts does have some evidence backing it: we do engage in various forms of indirect advising. There are three problems with both the directive and assertive versions of this view. First, as we observed in section 3.2.1, it is a substantive claim that all assertive advising or all directive advising involves indirect speech acts. We should posit indirect speech acts only where we have good evidence for their existence, and there seem to be many cases of assertive and directive advising that do not involve indirect speech acts. An error theory could account for these cases, but that approach would also involve ascribing error to sociolinguists (see footnote 12). The second problem is that the linguistic evidence equally supports both the view that advisory assertives and askings involve indirect recommendations (the indirect directive account), and the view that advisory directives and askings involve indirect assertions (the indirect assertive account). If both views can appeal to similar kinds of evidence, any attempt to arbitrate between them will be futile, and both will collapse into the ambiguity view. Third, neither view is in a position to vindicate the function of the concept of advising. If the concept of advising is a tool to facilitate the pooling of deliberative resources, and deliberators have different kinds of needs, the choice of one kind of deliberative problem as central will be arbitrary.

The alternative is that advising is simply not a kind of speech act. Advising is an activity which we typically pursue via the means of speech, but it is not a kind of speech act. It is a kind of joint practical thinking. Rather than sifting through various complex indirect forms of advising to determine what advising is, we should see the variety of advising as illustrative of the diversity of forms which collaborative deliberation can take. For the purposes of typologies of speech acts, we might still think about the speech acts of advising-to, advising-that, and adviceless advising, and think about advising as a kind of speech act which involves moves within collaborative deliberation. The point remains that while the tools of philosophy of language are useful for mapping out the complexity of advising; we need to look to the theory of joint practical deliberation to understand what unifies it.

4. Conclusion

We started by considering what the point of our talk and thought about advising might be, framing the hypothesis that our concept of advising answers to our need to pool deliberative resources. The pooling of deliberative resources turns out to be a surprisingly complex activity, involving the provision of propositions, directions, and questions. This complexity is reflected in the diversity of forms which advising can take. Recognizing the diversity of advising is an important advance in our understanding, but we’ve also seen how to unify these diverse forms using the idea that advising is collaborative deliberation. This view also explains the possibility of adviceless advising, the distinctive modal force involved in advising, the relation between advising and prior normative facts, and why advisors need not be wise. I don’t think that we need to throw out previous work on advising in ethics and philosophy of language, but quite a bit of it turns out to be about special cases of advising.

Seeing advising as collaborative deliberation can help us get better at advising and at being advised. By understanding the different needs of deliberators and the diversity of forms which advising can take, we can better tailor our style of advising to the needs of advisees. And by understanding collaborative deliberation, we can understand when as advisor is overstepping their responsibilities and advising is coming close to venting, joint deliberation, or ordering.

In closing, I want to note three issues for future research.

There is no reason to think that advising is the only activity that cross-cuts the categories proposed by typologists of speech acts. It might well be that other communicative verbs which take both declarative and infinitival clauses and systematically bridge grammatical moods—“promise”, “tell”, “warn”, “guarantee”—will turn out to not be kinds of speech acts, but rather kinds of joint thinking or joint action which we pursue via the means of speech acts.30 One might worry that much of the evidence that suggests that advising is not a kind of speech act will overgenerate, committing us to the view that there are very few (if any) kinds of speech act. This worry is worth taking seriously, but to my knowledge there are no communicative verbs that are connected with declaratives, imperatives, and interrogatives as systematically as “advise” is.

We have been focusing on advising in general, and the majority of high-stakes advising—think of government advisors, financial advisors, and lawyers—involves professionals with role responsibilities which shape and limit the way in which they can advise, and place responsibilities which make offering certain kinds of advice non-discretionary. It would be interesting to try to understand what the norms of professional advising are, and how they might shape distinctive forms of collaborative deliberation.

There is a rich connection between advising and friendship: we evaluate friends by the quality and quantity of their advising, and seeking out advice is a way to deepen that friendship. If advising involves treating someone else’s practical situation as your own, then advising will have connections to the Aristotelian ideal of friendship as treating someone as an other self.

Notes

  1. See (Nowell-Smith 1954), (Gauthier 1963), (Stewart 1978), (Wiland 2000b, 2021), and (Hinchman 2005).
  2. I use “she” for advisors, “he” for advisors, and singular “they” for incidental characters.
  3. The French conseiller appears anomalous without a complement clause specifying advice. (1) is odd and perhaps ungrammatical:
    • (1) ?Alix m’a conseillé.

    Interestingly, some cleft constructions appear acceptable:
    • (2) C’est Alix qui m’a conseillée.

    There are two possible explanations for the difference: (i) the cleft construction triggers ellipsis, meaning that (2) is not referring to adviceless advising; (ii) the default information structure for advising reports includes what was advised, making (1) bad because it does not provide the expected information, whereas the cleft construction in (2) foregrounds the question of who advised me, which (2) provides. The first explanation explains away the appearance of adviceless advising, while the second gives a pragmatic explanation for the badness of talking about adviceless advising (if the default question is who advised whom to do what?, a sentence like (1) will be pragmatically infelicitous and not syntactically bad). An informal survey of Dutch, Greek, and German suggests that it is acceptable but awkward to use cognates of “advise” without a content clause.
  4. On the possibility of advising without speech acts, see (Searle 1979, pp. 6–7).
  5. We should distinguish between the function of a concept, the function of the thing, and the function of the speech acts used in talking about the thing. Following (Habgood-Coote 2019) I take functional approaches to focus on the functions of concepts, reflecting the fact that our concepts are answerable to our collective practical needs.
  6. See the Joey/Ross storyline in the Friends episode The One with the Inappropriate Sister.
  7. Wiland draws on legal cases to argue that, in some cases, advisor and advisee are engaged in a form of shared activity (Wiland 2021, chapter 7). Although we might talk about advising in connection to shared activity, I take the Tariq/Hannah case as evidence that this is a fringe use that ought not be central to how we understand advising. In these cases, the language of advising might mask what in reality are commands, joint decisions, or threats.
  8. The dinner case emphasizes that collaborative deliberation occurs within other deliberative contexts. If you and I are making a cake together, and I have the job of removing the cake from the oven, I can ask you for advice about whether the cake is done, since the question of when to take it out is my responsibility. We can also embed collaborative deliberation within pretenses. In a job interview, an interviewer might ask for hypothetical advice. Here the interviewee is asked to engage in collaborative deliberation within a pretense, so we might think of the response as a pretense of a pretense of joint thinking.
  9. The translations for “advise that” are of doubtful acceptability in French, German, Dutch, and Italian. In Greek simvulevo only takes declaratives when they involve priority modals (Oikonomou 2021).
  10. We can report imperatives using declaratives: “go to the shops!” can be reported by “Jared said that I ought to go to the shops.” Does this mean that all cases of advising-to are cases of advising-that? Speech act reports are notoriously flexible: here the reporter seems to confuse what the advisor has said with a modal statement made true by what they have said, meaning that the report is loose speech. An analogy: one might loosely report a declarative sentence using an epistemic modal: if John says“the shop is open”, this might be reported (loosely) by “John said that the shop ought to be open.”
  11. Nowell-Smith (1954, chapter 11) distinguishes between advising involving what he calls aptness-words—such as “the film is entertaining”—and gerundive-words—such as “the film is worth seeing”—suggesting that the former merely contextually entails a recommendation, while the latter explicitly commends a course of action. Gauthier (1963, pp. 50, 53–5) and Hamblin (1987, p. 11) point out that advising can take the form of either imperatives or declaratives, and that the latter need not recommend a course of action. Searle is the first to note that advising can have the illocutionary point of both assertives and directives, proposing the distinction between advising-that and advising-to (Searle 1979, pp. 28–9; see also Stewart 1978, p. 204, and Raz 1979, fn14).
  12. Heritage and Sefi (1992) studied health visitors working with first-time mothers. They distinguish between advising involving recommendations, imperatives, deontic modals, and factual generalisations (1992, pp. 368–9). In a study of district nurses, Leppänen (1998, p. 223) classified recommendations given using imperatives, deontic modals, presentations of proposed actions as alternatives, and descriptions of future actions as advising. Locher’s typology of advising includes declaratives, questions, imperatives, referrals to other experts, general information, descriptions of one’s own experience, explanation, and metacommentary (Locher 2006, pp. 63–69).
  13. See (Locher 2006, p. 65) for examples. Sartre’s famous response to his student, “vous êtes libre, choisissez, c’est-à-dire inventez” [you are free; choose! Which is to say invent!] (Sartre 1946, p. 47, my translation) also appears to be a case of adviceless advising, despite the use of imperatives.
  14. Silverman et al. (1992) studied HIV counselling in England and the United States, finding both information-delivery and interview styles of advising. They found that counsellors often switched between these styles, but the authors give many examples of discourse fragments where counsellors only asked questions (1992, pp. 75–78). In her corpus of 280 online advice columns, Locher classified 31% of advisory moves as imperatives inviting future action, 5% as imperatives inviting introspection, 2% as interrogatives inviting actions, 9% as interrogatives inviting introspection, and 52% as declaratives (Locher 2006, p. 88). Although her category of interrogatives inviting actions plausibly involves indirect direction, the category of interrogatives inviting introspection is non-directive, except in the sense that asking any question is a proposal to answer it.
  15. (Stewart 1978, p. 207, fn17).
  16. I am not suggesting that following advice is morally deficient (see Hills 2009), just that we should care both about doing the right thing and about being able to work out what to do.
  17. This paragraph takes inspiration from Craig’s discussion of knowledge-how (Craig 1990, chapter 17).
  18. I take this terminology from (Portner 2007).
  19. A caveat: if the advisee’s problem involves shared reasons or role-responsibilities, we shall find a broader range of modals in advising.
  20. The possibility of persuasion in the context of advising means that—contra (Andreou 2006)—advising does not provide support for motivational internalism.
  21. The acceptability of persuasion is contextual and may be limited by role responsibilities. It would be surprising for a financial advisor to try to persuade you to care more about your family.
  22. For disagreement about the significance of this distinction, see (Raz 1979), (Hamblin 1987, pp. 10–14), (Wiland 2000b, 2004), (Darwall 2006, pp. 12–13), (McMyler 2011, chapter 5). We need to qualify this claim: sometimes the point of imperatival advice is to break deadlocks. Think about advice given to an advisee who faces a Buridan’s-ass situation with multiple equally good options, or advice given to a group facing a coordination problem with multiple equilibria.
  23. A general observation: all types of speech acts can modify the commitments of speaker, hearer, or both. In the case of assertions, unmarked declaratives (“It’s cold out”) propose shared commitment, rising declaratives (“It’s cold out[↑]”) propose hearer commitment, while falling declaratives (“It’s cold out [↓]”) propose speaker commitment (Gunlogson 2004, Portner 2018, pp. 207–8). For directives, we can propose hearer commitments (“feed the cat!”), commitments for both speaker and hearer (“let’s feed the cat!”), or commitments for the speaker (“let me feed the cat!”). Interrogatives can also generate commitments for the speaker, as with self-directed questions in essays. On the relation between speech acts and commitments, see (Geurts 2019).
  24. The question of what direction of fit askings involve is complicated: the presuppositions of a question represent the world, as do its answers, but asking a question plausibly involves the proposal to answer it.
  25. For the view that imperatives express propositions, see (Kaufman 2012). For the view that interrogatives express propositions, see (Karttunen 1977).
  26. A different kind of Moore-style sentence for advising appears in Nowell-Smith (1954, p. 154), who suggests:
    • (1) You ought to climb it [the mountain], but I don’t advise you to.

    Gauthier concurs that this sentence is weird suggesting that in this sentence ”the speaker is advising both for and against in the same breath“ (1963, p153).
  27. To support the idea that some assertions do not involve indirect recommendations, consider:
    • (1) I advise you that the train is leaving in ten minutes, but I wouldn’t take it.

    This sentence is not odd, although the following one is:
    • (2) #I advise you to take the train in ten minutes, but I wouldn’t take it if I were you.

  28. Weak imperatives (“take an Oyster”) used to make suggestions or change option spaces do not determine normative facts in this way, either for ordering or advising.
  29. For example, the sentence
    • (1) Jane advised Harry that he was going to be late, and to take a taxi.

    is quite acceptable.
  30. See (de Kenessey 2020, 2023).

Acknowledgments

Thanks to Solmu Antilla, Lisa Bastain, Sebastian Becker, Jessica Brown, Thomas Brouwer, Étienne Brown, Jade Fletcher, Eleanor Gordon-Smith, Simon Graf, Daniel Harris, Sandra Lapointe, Brendan de Kenessey, Deborah Marber, Nerri Marsilli, Fabienne Martin, Jared Millson, Andrew Peet, Caroline Torpe Touborg, Michael Traynor, Laura Whateley, Robbie Williams, Dennis Whitcomb, Bill Wringe, three referees for this journal, and audiences at both the St Andrews Friday Seminar and the Katherine Hawley Memorial Symposium.

I wrote the first draft of this paper with Katherine Hawley in 2013–14, and both it and I owe an enormous debt to her.

This research received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement no. 818633).

References

Andreou, C. (2006). Standards, Advice, and Practical Reason. Journal of Moral Philosophy 3 (1): 57–67.

Anscombe, G. E. M. (1957). Intention. Harvard University Press.

Austin J. L. (1975). How to Do Things With Words. Harvard University Press.

Bach, K. and Harnish, R. (1979). Linguistic Communication and Speech Acts. MIT Press.

Belnap, N. (1990). Declaratives Are Not Enough. Philosophical Studies 59(1): 1–30.

Bratman, M. (1992) Shared Co-operative Activity. The Philosophical Review 101(2): 327–341.

Craig, E. (1990). Knowledge and the State of Nature: An Essay in Conceptual Synthesis. Oxford University Press.

Darwall, Stephen L. (2006). The Second Person Standpoint: Morality, Respect, and Accountability. Harvard University Press.

Diederich, C. and Höhn, N. (2012) “Well it’s not for me to advise you, of course…”: Advice and Advise in the British National Corpus of English. In Miriam Locher and Holger Limberg (eds.), Advice in Discourse. John Benjamins: 333–358.

Gauthier, D. (1963). Practical Reasoning. Clarendon Press.

Geurts, B. (2019). Communication as Commitment Sharing: Speech Acts, Implicatures, Common Ground. Theoretical Linguistics 45(1–2): 1–30.

Gibbard, A. (2003). Thinking How to Live. Harvard University Press.

Gilbert, M. (2009) Shared Intention and Personal Intentions. Philosophical Studies 144(10): 167–187.

Groenendijk, J. and Stokhof, M. (1984). On the Semantics of Questions and the Pragmatics of Answers. In Fred Landman and Frank Veltman (eds.), Varieties of Formal Semantics: Proceedings of the Fourth Amsterdam Colloquium. Foris: 143–170.

Gunlogson, C. (2004). True to Form: Rising and Falling Declaratives as Questions in English. Routledge.

Habgood-Coote, J. (2019). What’s the Point of Knowing How? European Journal of Philosophy 27(3): 693–708.

Hamblin, C. (1987). Imperatives. Basil Blackwell.

Harris, D. (2014). Speech-Act Theoretic Semantics. Dissertation, City University of New York.

Harris, D. (2021). Imperative Inference and Practical Rationality. Philosophical Studies 179(4): 1065–90.

Heritage, J. and Sefi, S. (1992). Dilemmas of Advice: Aspects of the Delivery and Reception of Advice in Interactions between Health Visitors and First-Time Mothers. In P. Drew, & J. Heritage (Eds.), Talk at Work Cambridge: Cambridge University Press: 359–419.

Hills, A. (2009). Moral Testimony and Moral Epistemology. Ethics 120(1): 94–127.

Hinchman, E. (2005). Advising as Inviting to Trust. Canadian Journal of Philosophy 35(3): 355–385.

Hobbes, T. (1998). On the Citizen. Cambridge University Press.

Hobbes, T. (2012). Leviathan. Clarendon Press.

Karttunen, L. (1977). Syntax and Semantics of Questions. Linguistics and Philosophy 1 (1): 3–44.

de Kenessey, B. (2020). Promises as Proposals in Joint Practical Deliberation. Noûs 54(1): 204–232.

de Kenessey, B. (2022). Promises, Offers, Requests, Agreements. Ergo 9(34): 878–910.

Leppänen, V. (1998). The Straightforwardness of Advice: Advice-Giving in Interactions between Swedish District Nurses and Patients. Research on Language and Social Interaction 31(2): 209–239.

Locher, M. (2006) Advice Online: Advice-Giving in an American Internet Health Column. John Benjamins.

McMyler, B. (2011). Testimony, Trust, and Authority. Oxford University Press.

Nagel, J. (2019). Epistemic Territory. Proceedings and Addresses of the American Philosophical Association 93: 67–86.

Nowell-Smith, P. H. (1954) Ethics. Penguin.

Oikonomou, D. (2021). Modally Conditioned Mood-Switch: The Case of Advise-Predicates in Greek. Proceedings of Semantics and Linguistic Theory (SALT) 31: 662–682.

Portner, P. (2007). Imperatives and Modals. Natural Language Semantics 15(4): 351–383.

Portner, P. (2018) Commitment to Priorities. In: D. Fogal, D. Harris, and M. Moss (eds.), New Work on Speech Acts. Oxford University Press: 296–316.

Raz, J. (1979) Legitimate Authority. In The Authority of Law: Essays on Law and Morality. Clarendon Press, 2–27.

Roberts, C. (1996/2012). Information Structure in Discourse: Towards an Integrated Formal Theory of Pragmatics. Semantics and Pragmatics 5: 1–69.

Roberts, C. (2018) Speech Acts in Discourse Context. In D. Fogal, D. Harris, and M. Moss (eds.), New Work on Speech Acts. Oxford University Press: 317–359.

Sartre, J. P. (1946). L’Existentialisme est un Humanisme. Nagel.

Searle, J. (1969). Speech Acts: An Essay in the Philosophy of Language. Cambridge University Press.

Searle, J. (1979). A Taxonomy of Illocutionary Acts. In Expression and Meaning, Cambridge University Press, 1–29.

Searle, J. and Vanderveken, D. (1985). Foundations of Illocutionary Logic. Cambridge University Press.

Silverman, A. and Bor, R. (1992). Discussing Safer Sex in HIV Counselling: Assessing Three Communication Formats. AIDS Care 4(1): 69–82.

Sliwa, P. (2012). In Defence of Moral Testimony. Philosophical Studies 158(2): 175–195.

Stewart, D. (1978). An Analysis of “Advising.” Educational Theory 28(3): 202–213.

Vendler, Z. (1972). Res Cogitans: An Essay in Rational Psychology. Cornell University Press.

Wiland, E. (2000a). Good Advice and Rational Action. Philosophy and Phenomenological Research 60(3): 561–569.

Wiland, E. (2000b). Advice and Moral Objectivity. Philosophical Papers 29(1): 1–19.

Wiland, E. (2003). Some Advice for Moral Psychologists. Pacific Philosophical Quarterly 84(3): 299–310.

Wiland, E. (2004). Trusting Advice and Weakness of Will. Social Theory and Practice 30(3): 371–389.

Wiland, E. (2021). Guided by Voices: Moral Testimony, Advice, and Forging a “We.” Oxford University Press.

Zwicky, A. and Sadock, J. (1975) Ambiguity Tests and How to Fail Them. Syntax and Semantics 4, 1–36.