Hope and anxiety are familiar and ubiquitous characters of our affective lives. Insofar as we care about the future, we are creatures of hope and anxiety. For creatures like us, entertaining possible future scenarios of how our life might play out – where we will be, what could happen – is often accompanied or “charged” with emotions like hope and anxiety. We manifest our concern for our personal future in part through our disposition to experience anxiety and hope about relevant possibilities. As J.P. Day remarks: “the farmer hopes that it will rain tomorrow […] the hostess fears that it will rain tomorrow […]. Only the dispassionate meteorologist believes that it will or will not rain tomorrow” (1998, p. 122). Depending on our goals, concerns, and desires, we have hopes and anxieties about mundane everyday events like the weather, but also more substantial hope and anxiety about the direction that our life as a whole is taking. We get anxious about important life choices we are about to make, and we hope to achieve our most valued goals.

What will interest me in this article is whether the epistemic profile of hope and anxiety, and in particular the fact that they are directed at uncertain outcomes, might pose a threat to the stability of their valence. It is widely agreed that all affective states share two dimensions of valence and arousal (Reisenzein & Spielhofer, 1994; Russell, 1980, 2003). The valence of an emotion refers to its positive or negative ‘charge’ or force. It is, as some have famously called it, the “heat of emotion” (Charland, 2005). It is commonly assumed that emotions and their associated feelings can be defined as either positive or negative. It is even thought that valence as an essential property of emotions is a promising criteria for demarcating emotional states from cognitive states (Charland, 2005). The idea that emotion types (e.g. anger, sadness, admiration, pride, etc.) are divided between those that are positive and those that are negative, is well established in the literature (Colombetti, 2005). However, the nature of emotional valence itself and the question of where emotions draw their valence from is still a matter of important debate in the philosophy of emotion. Within this debate, certain emotions (such as the emotion of surprise or the feeling of sublime) are viewed as posing a challenge, as they do not seem to carry a clear negative or positive charge. The question that will interest us here is: might emotions of uncertainty like hope and anxiety also constitute counter-examples to the idea that emotions have a necessary binary valence?

Insofar as hope and anxiety are directed at states of affairs the occurrence of which is uncertain, these emotions seem to involve both an awareness of how events could unfold positively, and an awareness of how events could unfold negatively. Hope necessarily involves an awareness of the risk of non-attainment of the hoped-for outcome, and anxiety necessarily involves an awareness of the possibility of non-occurrence of the threatening outcome. As long as the future remains uncertain, we might hope for the best, and still be anxious that events do not unfold as we hope they will.Footnote 1

Our emotional engagement with future events often involves both a felt awareness that the future could unfold positively (yield positive outcomes), and a felt awareness that it could unfold negatively. For this reason, our affective apprehension of uncertain events often takes on a sweet-and-sour taste. This has led several philosophers to wonder whether hope and anxiety could really be strictly and consistently positive and negative emotions. It has been argued that, because hope and anxiety have uncertainty built into their intentional content, the valence of these emotions might not be as clear as we thought. Does the epistemic profile of hope and anxiety pose a threat to the stability of their valence? Is the valence of hope and anxiety then more blurry, or mixed than what we previously thought? Could this explain the often ambivalent nature of our emotional engagement with future or uncertain events?

In the paper I address these questions as follows. In section I, I explain what it means for hope and anxiety to belong to the small category of knowledge-precluding or uncertainty emotions (Gordon, 1969), and what this entails with regard to the nature of their intentional objects. In section II, I introduce the widespread idea that each emotion type carries a fixed valence, and I put forward some initial intuitive reasons to doubt that this principle applies to hope and anxiety. In section III, I present a first account according to which anxiety is not in fact a negative emotion. According to this view, anxiety does not have a strictly negative valence, but is instead a mixed emotion composed of conflicting attitudes (Miceli & Castelfranchi, 2015). I will show that this account of anxiety does not hold, and argue instead that anxiety is a sui generis emotion with a modal object. In section IV I present a recent account of hope which also challenges the emotional valence principle, and according to which hope is not strictly a positive emotion, as some specific occurrences of hope are in fact negatively valenced (Stockdale, 2019). I will object to this account of hope by arguing that the phenomena the author is referring to are in fact not genuine episodes of hope. Finally, in section V, I explain where I believe the doubt regarding the valence of these emotions stems from. Moreover, I argue that, in spite of the folk psychological idea, supported by some philosophical accounts, according to which hope and anxiety are not strictly and consistently valenced, we have a good reason to think that these emotions are instead respectively, and necessarily, positive and negative.

1 Hope and Anxiety: Knowledge-Precluding Emotions

I want to start by clarifying an important terminological point. While philosophers who have discussed the relationship between hope and anxiety have typically used the term “fear” as opposed to the term “anxiety” within those debates, I believe it is clear given the claims, assumptions, and paradigmatic cases they present, that the phenomenon they have in mind is anxiety.

While anxiety has long been ignored by philosophers, the literature in the philosophy and psychology of emotion has now clearly acknowledged it as as an emotion distinct from fear. Fear and anxiety are roughly distinguished in the following way: while the former is a response to a danger that is appraised as present, the latter is a response to a danger that is appraised as only possible (Kurth, 2015, 2018). More precisely, anxiety is focused on the uncertain features surrounding a threat, and specifically the uncertainty regarding its occurrence (whether it will, in fact, materialize or not). For instance, if I feel anxious about taking the plane tomorrow, I apprehend taking the plane as implying a possible threat, usually: a plane crash. However, the intentional object of my anxiety is not the danger per se; it is not the plane crash. That is the intentional object of fear. In anxiety, the focus is rather on the uncertainty surrounding the threat: whether it will in fact occur.Footnote 2 The intentional object of fear is a possible danger, whereas the intentional object of anxiety is a possible and uncertain danger.Footnote 3 In other words, anxiety involves the apprehension of potential negative outcomes (implied by some particular event or situation) over which we lack information. In a more concise manner, Charlie Kurth (2015, p. 5) has proposed to formulate the intentional object of anxiety as a “problematic uncertainty”. For the sake of accuracy, I will here be using the term “anxiety” where many authors have used the term “fear”, to refer to short-term emotional episodes directed at problematic uncertainties or uncertain threats.

The relationship between anxiety and hope has interested such eminent philosophers as Spinoza, Hume, and Descartes. As Hume has remarked, the reason why we might feel both of these emotions when thinking about possible future outcomes, is that both emotions have uncertain states of affairs as their intentional objects. Both emotions are directed at states of affairs which one does not know, or take oneself to know to be true or false. Feeling hope or anxiety that p requires the absence of knowledge regarding whether p will occur.

Hope and anxiety have a distinctive epistemic profile. Both emotions are said to belong to the small family of uncertainty or knowledge-precluding emotions (Gordon, 1969, 1987). Hoping that p or being anxious that p is inconsistent with knowing whether p is or will be the case. In other words, there seems to be a psychological incompatibility between feeling anxious or hopeful and being certain that it will happen. This is manifest, for instance, in the fact that statements involving the self-ascription of hope or anxiety (1) (2), followed or preceded either by an assertion of the fact itself, or a self-ascription of a factive emotion (3) (4), are infelicitous:

  1. 1)

    The cyclist fell. I hope that she doesn’t fall.

  2. 2)

    The cyclist fell. I am anxious about her falling.

  3. 3)

    I regret that the cyclist fell, but I hope that she doesn’t fall.

  4. 4)

    I regret that the cyclist fell, but I am anxious that she might fall.

One can feel these emotions only when one does not know that p, and these emotions vanish as soon as one comes to know p. As Benton (2019, p, 1) notes:

when one comes to know that what one hoped for obtains, one’s attitude changes from hope to satisfaction, or even joy, at learning that one’s hope was fulfilled. Yet when one comes to know that what one hoped for does not obtain, one’s attitude changes from hope to mere wish, or even regret, at learning that one’s hope was dashed. In both cases it would be somehow irrational, perhaps even psychologically impossible, to retain one’s hope upon coming to know the outcome.

In the same way, when one comes to know that the threat one was anxious about has occurred or will occur for certain, one’s attitude changes from anxiety to dread, despair, or resignation, and it would be somehow irrational, and perhaps psychologically impossible, to keep being anxious about something once one learns that it has already occurred or that it is for certain occurring. Hope is thus importantly distinct from the emotions one feels when one apprehends a desired state as obtaining, and anxiety is distinct from the emotions one feels when one apprehends a menace as occurring.

The idea I will introduce next is that the fact that hope and anxiety have this particular epistemic profile might suggest that their valence is not as fixed and stable as we take it to be.

2 Uncertainty Challenges Emotional Valence

While hope and anxiety are both directed at uncertain states of affairs, hope is considered a “positive” emotion, and anxiety a “negative” emotion. In other words, it is commonly understood that what distinguishes anxiety and hope is their valence.

Valence is thought to be an essential attribute of emotions. It is generally assumed that types of emotions, such as guilt, pride, sadness, etc., are divided between those that have a negative valence, and those that have a positive valence. This is reflected in our claims that, for instance, anger is negative, and pride is positive. Each emotion type is thought to have a valence that is either positive or negative, and that is consistent across all occurrences of that emotion. Here I will indeed be interested in “emotional valence”, that is, valence as a property of whole emotions, as opposed to a property of individual bouts of affect, appraisals, or behaviours (Colombetti, 2005).

The fact that emotions inherently carry a negative or a positive force is thought to be what imbues them with personal meaning or concern (Frijda, 1986; Lazarus, 1991). However, the nature of emotional valence is still a matter of great debate in the philosophy of emotion. Determining which property of the emotions provides them with their valence is far from obvious (Colombetti, 2005). In this paper I will presuppose an attractive and often defended view of emotional valence, which I present in this section. This view, as we will see, is motivated by the observation that valence is a fundamentally evaluative notion, that is intimately linked to the motivational nature of emotions. Without valence and its accompanying ‘charge’ or ‘force’, emotions would not “impel us to act” (Prinz, 2004, p. 128) in the specific normative direction that they do.

However, certain emotions, such as surprise and the feeling of sublime for instance, have been presented as counter-examples to the principle that each emotion is either positive or negative (Dokic & Arcangeli, 2020; Noordewier & Breugelmans, 2013). I will here examine whether we have grounds to believe that hope and anxiety might also have a valence that is more complex than what we tend to think. This intuition can already be found in folk psychological notions, and in certain expressions we commonly use to refer to the complexity and ambivalence of our emotional engagement with the future. For instance, in our everyday language, we might speak of an episode of hope as infused with” or “tainted by” anxiety. Certain philosophers have suggested that hope does not have a fixed valence. Sabine Döring (2014, p. 124) has stated: “Sometimes hope is joyful, sometimes it is anxious; in any case it is always inspired by a feeling of confidence of a certain degree”.

Other philosophers claim that hope and anxiety may affect each other’s valence, due to the fact that our emotional engagement with uncertain events is often constituted by a mixture of both emotions. Spinoza argues that one cannot hope for an outcome without also fearing that it does not occur (Ethics III.P50). A few contemporary philosophers have continued to acknowledge the tight relationship between hope and anxiety. Averill claims that hope always involves anxiety, in the sense that: “a person cannot hope for something unless he or she also fears that the hoped-for event might not happen. Hope and fear are two sides of the same coin” (Averill, 1996, p. 36). This echoes Bovens’ claim that whenever one cares for the well-being of a loved one, one necessarily both hopes that the person be well, and fears that they won’t. Bovens (1999) states that: “hoping and fearing for the well-being of a loved one are constitutive of loving”.Footnote 4 How should we read these philosophical claims about hope and anxiety? And how should we understand these folk psychological expressions? What reasons do we have to doubt the valence of hope and anxiety? In what follows I will attempt to lay out some initial intuitive reasons, before presenting two philosophical accounts of hope and anxiety as ambivalent and unstably valenced.

As we have begun to see, hope and anxiety are different from other emotions. Hope and anxiety are not based on a belief; they are not based on a “settled” epistemic state, but on an “unsettled” epistemic state. The fact that both emotions are incompatible with one’s knowing whether or not the target state of affairs will come true distinguishes hope and anxiety from a large variety of emotions which instead require that one knows (or take oneself to know) that an event is happening or has happened. For instance, feeling sad (in an appropriate manner) requires one to believe that a loss has happened; feeling angry requires one to believe that an offense has been committed; feeling guilt requires that one believes that one has performed a wrong deed, etc. Emotions like anger, sadness, or guilt require that one takes an object or situation to be present or actual. In other words, they require that one believes the relevant facts about the particular objects of one’s emotion.

By contrast, the target objects of hope and anxiety are states of affairs that one takes to be uncertain. Hope is about a world in which I win the upcoming race, or pass my law degree next semester. Anxiety is about a world in which my medical exams show a complication, or the plane I embark on crashes, or my partner leaves me. All of these propositions express possibilities, states of the world which I take to be uncertain (that is, neither certain nor impossible). When I am sad about my bike having been stolen, I take it that it is the case that my bike has been stolen. By contrast, when I am anxious about my bike having been stolen, although I doubt that it might be the case, I do not take myself to know this. For all I know, it might or might not have been stolen. In other words, the question of the occurrence of their target event is one that is necessarily open and unresolved for the person who hopes or is anxious. How might this affect the stability of the valence of these emotions? To answer this question, we first need to say more about the nature of emotions, and about the property which emotions plausibly derive their valence from.

Philosophers commonly think that emotions are mental states which relate us to values (De Sousa, 1987; Deonna & Teroni, 2012; Goldie, 2002; Tappolet, 2016). Some philosophers believe that emotions relate us to values in virtue of being a kind of perception, or a kind of judgement, and others believe that emotions constitute sui generis evaluative attitudes which are not reducible to a kind of perception or judgement. While these theories differ in their accounts of how emotions relate us to values, they all conceive of emotions as involving a form of evaluation. Emotions are thus thought to have a twofold intentionality: they are directed at a particular object (a given situation, an event, etc.), and they apprehend this particular object as instantiating a certain evaluative property. The specific evaluative property that an emotion type makes salient or relevant in our environment or current situation is also called its formal object (Teroni, 2007). The formal object of an emotion refers to the property which the emotion as a type (i.e. “fear” or “sadness” for instance) generally tracks or responds to. Fearsomeness, funniness, and offensiveness are respectively the formal objects of fear, amusement, and anger. Each emotion type can thus be considered as a distinct type of affective evaluation in virtue of the fact that it targets a distinct formal object, i.e. a distinct evaluative property in its particular objects. Because it is the nature of emotions to relate us to evaluative properties in our environment, emotions constitute a kind of engagement with the world that is by definition not neutral. But why should emotions necessarily be bivalent states? Why should there be intrinsically positive and intrinsically negative emotions?

Valence plays a crucial psychological role in motivating intentional action, by enabling us to compare alternative courses of action on the basis of a swift signal generated by our evaluative systems (Gilbert & Wilson, 2005). Following a hedonic approach, it is widely accepted that emotions can be classified as positive or negative in virtue of their phenomenological feature, or “what it is like” to experience them (Lambie & Marcel, 2002). Carefully examining the phenomenology of some emotions, however, it is often not clear whether they are to be primarily associated with pleasure or displeasure (think of anger or surprise, for instance). Alternative attempts to classify emotions as positive or negative include considering 1) their motivational tendencies: attraction towards the object of the emotion, or aversion from it (McLean, 2008), (2) the desires and wants (to continue or stop) accompanying the emotion (Prinz, 2004), 3) the degree to which the emotion is congruent or incongruent with the subject’s goals (Lazarus, 1991).

As Carruthers points out, there are however few views of valence capable of accounting for the essential role of valence in motivation (Carruthers, 2018). The one I will favor here centrally appeals to the evaluative nature of valence. The valence of emotions, on this account, is explained by the fact that emotions are or involve the evaluation of an object as good in some particular manner, or bad in some particular manner. Indeed, it is commonly assumed that each type of emotion (such as anger, fear, sadness, etc.) is an experience of a thick valueFootnote 5 being instantiated in particular objects or situations. When we feel an emotion, what we experience is the value of that object or situation: in fear we experience the fearsomeness of the snake, in amusement we experience the funniness of the joke, in anger we experience the offensiveness of the remark (Deonna & Teroni, 2015; Tappolet, 2016). Crucially, thick values have polarity: they are either negative (the shameful, dangerous, disgusting, etc.), or positive (the funny, admirable, successful, etc.). Negative thick values are specific ways in which something can be bad (because it is shameful, or disgusting, or dangerous), and positive thick values are specific ways in which something can be good (because it is admirable, funny, or successful). And thus, the property able to explain the valence of emotions can be found in the fact that emotions are essentially defined as states which relate us to those values (Teroni, 2018). Given that values are polarized entities – they are necessarily either negative or positive the valence of an emotion type depends on whether this emotion type has a positive or a negative thick value as its formal object.

The account can be understood in various ways, depending for instance on whether one views emotions as involving a nonconceptual perception-like representation of value, or whether one views them as nonrepresentational states. For example, fear may be seen as representing the fast approaching vehicle as fearsome, or as otherwise apprehending it as fearsome. In any case, in this view it is the fact that an emotion amounts to the experiencing of an object, situation, sensation, etc. as in some way bad, which accounts for emotional valence. On this account, the reason why valence has a psychologically fundamental role in motivating intentional action is that it presents options and situations as good, bad, better, or worse to us; that is, as more or less positively or negatively valuable (Carruthers, 2018).

The fact that emotions are experiences of value explains why emotions are associated with the hedonic states that they are. What is unpleasant or pleasant in an emotional episode of some type is the fact that it relates us to some (positive or negative) value (funniness, fearsomeness, offense, loss, etc.). Emotions draw their pleasantness and unpleasantness from the values of which they are experiences: it is because pride is essentially an experience of a positive value (success) that it has a positive valence, and it is because sadness is an experience of a negative value (loss) that it feels unpleasant (Teroni, 2018). It also explains why emotions are associated with the behavioral orientations that they are: emotions move us in the way they do because they relate us to the values (danger, offense, etc.) that they do. In sum, the valence of an emotion is constituted by the fact that this emotion either consists in an experience of a positive thick value, or of a negative thick value.

Now remember that, by contrast with emotions like pride, anger, or guilt, in the case of hope and anxiety, the values (of some event as threatening or desirable) are experienced as instantiated by states of affairs which are uncertain. Hope and anxiety are not emotions which relate us to evaluative properties of actual events, they relate us to evaluative properties of relevant possibilities (possibilities which are meaningful to us, given our desires, goals, and concerns). While anger, sadness, and guilt are ways for us to apprehend evaluative properties as being instantiated in particular objects, hope and anxiety are ways for us to apprehend that evaluative properties might be instantiated in some object. The scenarios we hope for or are anxious about are per definition scenarios which we have not seen unfold in front of our eyes. We are anxious or hopeful about outcomes which we imagine could occur. As such, hope and anxiety constitute experiences of uncertain value, as opposed to experiences of actual value.

In a similar manner, some responses to the famous paradox of horror or tragedy in aesthetics consist in claiming that fiction-directed negative emotions are in fact not experienced as negative. This, as some have claimed, is in part due to our awareness of the fictional nature of the events depicted in the artwork and at which our emotional reactions are directed. We are able to appreciate the negative emotional experiences provided by horror films and tragedies, since:

our first order affective reactions to negative, unpleasant representational content are not straightforwardly intrinsically negative, for their valence is partly dependent, amongst other things, on the awareness of non actuality (Todd, 2014, p. 21).

Our awareness of the non actuality of the events to which we react emotionally, Todd claims, impacts the valence of our emotional experience. If this is valid, then one important reason for doubting the valence of hope and anxiety is that these emotions do not take evaluative properties to have been instantiated in actuality. Hope and anxiety do not present us with an evaluation that a threat has occurred or that a desired event has come true. They are states which relate us to the possibility that these may occur.

I have here provided initial intuitive reasons why the idea that each emotion type has a fixed valence may apply well to those emotions that are based on a belief about how things are, and not so well to those emotions that are about unknown or uncertain states of affairs. In the coming sections I present two distinct ways in which this claim has been defended in the recent literature on emotion. According to the first account, anxiety is an intrinsically ambivalent emotion, as a result of involving an awareness of conflicting possibilities or ways in which the future might unfold (Miceli & Castelfranchi, 2015). According to the second, hope is not a strictly positive emotion, as certain specific occurrences of hope are negatively valenced (Stockdale, 2019). I introduce and discuss the first account in the next section.

3 Anxiety as an Ambivalent Emotion

A first way in which we may argue that emotions of uncertainty challenge the theory of fixed valence is by arguing that they necessarily carry an ambivalence in their phenomenology, as a result of involving an awareness of both the possibility that events could unfold favorably, and the possibility that they could unfold unfavorably. One such view has been defended about anxiety.

Philosopher and psychologist Miceli and Castelfranchi, who have worked extensively on emotions of uncertainty (2015), have argued that anxiety is a mixed emotion, composed of an alternation of fear and hope. Anxiety, they argue, is where hope and fear “coexist”. Because anxiety necessarily implies not knowing whether a threat will occur, it involves an oscillation between these two states: as far as we know, events might unfold in a negative or in a positive manner (that is, in a manner in which our goals are promoted or thwarted). Anxiety is thus constituted by these two opposing attitudes regarding how events might unfold. In their view, the phenomenology of anxiety testifies to the conflict between hope and fear that is constitutive of anxiety. The feeling of trepidation that accompanies anxiety is due to the fact that, when we feel anxious, our attention is alternatively drawn to considering each possibility and their consequences in turn.

What is central to the experience of anxiety is the uncertainty about whether a threat will materialize. As long as this uncertainty remains, they argue, we will inevitably oscillate between conflicting affective attitudes. They write (2015, p. 133):

The anxious state of mind implies both the belief that the outcome might be negative, and the opposing (albeit typically more feeble) belief that the outcome might be positive. That is, it implies the coexistence of a negative IAR (Interested Anticipatory Representation) and a positive one—a mixture of, and conflict between, fear and hope, which accounts for the state of uncertainty and restless wait that is typical of anxiety

In this account of anxiety, the absence of knowledge regarding whether p will occur, which is necessary to the experience of anxiety, leads us to sway between two conflicting views as to what we think we should expect.

In support of their account of anxiety, Miceli and Castelfranchi present clinical data regarding differences between patients with anxiety disorders and patients suffering from depression. Patients with generalized anxiety disorder (GAD) – a psychiatric condition characterized by an omnipresent feeling of anxiety – can be differentiated from patients with depression by the dimension of positive affect present in the former, but not in the latter group. Whereas negative affect is shared by both groups of patients, low levels of positive affect are specific only to depression (Brown, 1997; Clark & Watson, 1991). Moreover, according to studies which compare the typical “expectations” of anxious versus depressed individuals, it has been shown that anxious participants are likely to anticipate more future negative experiences, but not fewer positive ones. Depressed participants, by contrast, anticipate fewer positive experiences, but not more negative ones (MacLeod et al., 1997). It is as though, Miceli and Castelfranchi conclude, anxious individuals did not “give up” or reduce their positive anticipations in favor of their negative ones. They anticipate both positive and negative experiences, and typically oscillate between them.

Importantly for us, the authors suggest that, as a result of being composed of these conflicting attitudes, anxiety does not have a strictly negative valence. It instead has a mixed valence. What are mixed emotions and can we make sense of anxiety as carrying a mixed phenomenology? Mixed emotions are thought to be affective states involving a co-existence of negative and positive feelings in an individual at a given time. A common example of such phenomena is nostalgia. Nostalgia is both positive and negative: positive because it apprehends a situation as good for one, and negative because it apprehends this situation as being no longer the case (Prinz, 2004, p. 165). In other words, nostalgia feels both good and bad because it is composed of two mental episodes which occur at the same time: one of apprehending a time period as good, and one of apprehending a time period as lost. Miceli and Castelfranchi’s suggestion seems to be that we view anxiety in the same way: if each time we feel anxiety we both feel that events could unfold unfavorably, and that events could unfold favorably, anxiety seems to necessarily carry a mixed phenomenology, akin to that of nostalgia. In what follows I am going to argue that anxiety cannot be understood on the model of nostalgia, and that we have a good reason to deny that the states which are said to compose anxiety can even get mixed.

To conceive of anxiety as a mixed emotion is to conceive of the states which constitute anxiety as forming a mixture. Mixed feelings are feelings belonging to distinct mental episodes, which occur at the same exact time in an individual. Hence, there are constraints bearing on the kind of states which may get mixed with one another, because not all mental states can actually co-occur. One constraint is that feelings may co-occur as long as they are not directed at the same aspect of the same object. The feelings that compose nostalgia can co-occur because nostalgia involves two mental episodes directed at the same object (a past moment), considered under a different aspect. One of the mental state composing nostalgia evaluates this moment under the aspect of its goodness, and the other evaluates it under the temporal aspect of its pastness. By contrast, the states which Miceli and Castelfranchi take as composing anxiety present us with two diverging and incompatible views of how one and the same event could unfold: the possibility of harm, and the possibility of safety, or the possibility of victory, and the possibility of defeat.Footnote 6 Now, can a mental state representing an event as occurring get mixed with a mental state representing that event as not occurring?

A serious problem I see with this view regards the plausibility that one could psychologically both be in a state of hope and in a state of fear at the exact same time. It seems indeed that Miceli and Castelfranchi do not mean to say that fear and hope are alternating to give rise to the state of anxiety, but that they properly co-exist. There is first a doubt regarding whether we are the kind of creatures that are able to simultaneously entertain incompatible hypothetical scenarios in mind. Hypothetical thinking seems in fact to be constrained by what Evans (2019) calls the singularity principle, according to which we can only consider possible scenarios one at a time, in a sequential or serial manner. In other words, if fear and hope have distinct contents (good and bad scenarios), and if the contents of two mental states cannot be entertained mentally at the same time, then these states cannot get mixed (Massin, 2018).Footnote 7 Furthermore, fear and hope are not only distinguishable in virtue of their contents, but also in virtue of constituting distinct modes or ways of evaluating their content. Fear and hope can be viewed as constituting different types of evaluative reactions to uncertain events. Apprehending an event with hope consists in evaluating it as promising, encouraging (Milona & Stockdale, 2018), as meriting longing and pursuit, while apprehending an event with fear consists in evaluating it as dangerous, as meriting avoidance, aversion, protection. It is then difficult to conceive of a single mental state whereby one would simultaneously manifest these two opposite reactions of aversion and longing towards the very same object at the very same time. If we cannot embody opposite evaluative reactions to one and the same event at the same time, then the two distinct attitudes which are said to compose anxiety in this account cannot properly speaking get mixed.

The states of trepidation and restlessness that Miceli and Castelfranchi mention are nonetheless real. Upon waiting to see one’s name appear on the nominee’s list after the championship, one is likely to be emotionally unsettled. While Miceli and Castelfranchi argue that this is a typical manifestation of anxiety, I argue that it should be understood otherwise. Explaining these experiences does not require us to posit that anxiety is itself a mixed and composed state. These experiences of oscillation and emotional unsettledness can be explained as the typical manifestation of an alternation of anxiety with the emotion of hope, without the episode of hope being a constitutive element of the episode of anxiety. As emotions of uncertainty, hope and anxiety will tend to co-occur when faced with goal-relevant events implying uncertain outcomes. When they do, they will alternate in our stream of consciousness, having our attention move, sometimes very rapidly, from reacting (with anxiety) to the possibility that I might not be selected, to reacting (with hope) to the possibility that I might be selected for the competition.

In this sense, hope and anxiety tend to live on the backdrop of each other. This explains why our affective engagement with uncertain outcomes often seems to take a mixed, unstable shape. I believe we can now also easily see how the clinical data regarding GAD and depressive patients can be explained without having to posit that anxiety intrinsically involves positive affect. Given that anxiety has a modal object (“a threat might be upon us”), anxiety does not involve a belief about what will happen; it involves at best a suspicion, or a doubt. As such, experiencing anxiety does not preclude one from also experiencing hope. If this is so, there is no reason why anxious individuals could not also feel hope, which explains the presence of positive affect and negative affect in these individuals. The fact that anxious patients do experience positive affect does not mean that hope is a necessary element of anxiety, it simply means that feeling (even very) anxious does not preclude feeling hopeful. By contrast, the emotions that depressive individuals typically experience, such as sadness and despair, do preclude hope. One could even argue that despair is precisely defined by the absence of hope: as soon as one is disposed to hope, one does not despair anymore. Depressed individuals do not experience positive affect or positive expectations, because the emotions that define their overarching engagement with the world are emotions which preclude hope.

I have shown that the account of anxiety as a mixed emotion does not hold. What we are left with is the necessity for hope and anxiety to constitute two distinct mental episodes, which may at best alternate with each other, but cannot get mixed unless they are directed at distinct objects. In the next section, I introduce another way in which emotions of uncertainty have been argued to challenge the idea that each emotion type is either negative or positive. In this account, it is not that their valence is intrinsically mixed, it is that their valence is not consistent: it changes from one episode to another.

4 Negative Hopes

Another way to argue that emotions of uncertainty challenge the theory of valence is to claim that certain occurrences of these emotions do not carry the valence that we typically associate with them. It is the claim that valence is not a fixed property of an emotion type, but that it varies across different occurrences of a same emotion type. In a recent article, Katie Stockdale (2019) argues that some occurrences of hope are negatively valenced, and that hope is thus not a strictly positive emotion. The phenomenology of hope, she claims, is not always as bright and cheery as philosophers have assumed and described it to be.

More precisely, Stockdale (2019) has argued that hopes which are felt in the context of a menace are negatively valenced episodes of hope. They are what she calls “fearful hopes”: hopes to escape a menace, avoid a difficult situation, overcome obstacles, etc. The kinds of hopes which manifest in threatening contexts, Stockdale explains, are partly constituted by fear. In these cases, “fear actually constitutes hope and does not just exist alongside it” (Stockdale, 2019, p. 119). Because these episodes of hope are partly constituted by fear, they are not positively valenced like other episodes of hope.

Let us take a closer look at Stockdale’s characterization of the phenomenon of “fearful hopes”. One example she provides is that of a woman called Suzie who is walking home at night and is aggressively addressed by a man. At that moment, Suzie thinks to herself “I really hope this man leaves me alone.” Suzie envisages that the man might assault her, and she is hoping that he doesn’t. The second example Stockdale presents is that of a black woman who “fearfully hopes” that the police officer who is walking towards her vehicle does not use violence against her. Additional examples include: an indigent person who is “fearfully hoping that the water dripping from one’s ceiling is not a sign of serious, expensive damage” (Stockdale, 2019).

Upon reflecting on these particular cases of hope, Stockadle (2019, p. 121) remarks: “there is no sense in which women or people of color experience anything like pleasant feelings of anticipation in hoping to escape violence.” Indeed, when we hope to escape a menace, we mainly feel bad: we do not feel the pleasurable longing, and the behaviors of motivated pursuit typically associated with the experience of hope. In other words, some hopes seem to carry none of the typical markers of a positive emotion: they are not accompanied with pleasant feelings, or the desire to move towards a positive goal. “Fearful hopes” are instead accompanied by unpleasant feelings and behavioral tendencies of avoidance and withdrawal from the situation. Stockdale’s conclusion is that these occurrences of hope are negatively valenced. This, in her view, should invite us to reconsider the valence of the emotion of hope altogether: if certain types of hope feel bad, “hope is thus not positively valenced in the way that many philosophers have supposed hopes to be” (Stockdale, 2019, p. 121).

Now, we have to ask, do the cases provided by Stockdale challenge the nature of hope as a negative emotion? Do they then challenge the theory of valence according to which each emotion type is consistently either positive or negative? I think they do not. The reason for this is simply that “fearful hopes” as defined by Stockdale do not, in fact, qualify as episodes of hope. Stockade’s account of these phenomena trades on an ambiguity between two senses of “hope” in everyday language. In everyday language, we use the term “hope” to refer to two different psychological phenomena, only one of which is actually an occurrent emotional episode. In everyday language, we use “hope” to report an emotional episode or an occurrence of the emotion type of “hope”, and we also use it to express a desire that we have, for things to go one way or the other.

While we use “hope” to refer to both an emotion and a desire, these are two importantly distinct psychological phenomena. While an emotional episode occurs and appears in our stream of consciousness, a desire is a background disposition. Mental states which occur are distinct from mental states which merely dispose us to have certain experiences. Crucially, given that dispositions do not occur, they have no phenomenology. Stockdale’s claim, I argue, is based on a false assumption that reports such as “I really hope this man leaves me alone” indicate an underlying emotional episode of hope. I instead believe that these reports indicate a desire in Suzy, that the man leaves her alone.

Now, you might object that the experiences which Stockdale describes do have a phenomenology – they feel bad. How could it be, then, that these experiences are not emotional episodes of hope? This is because a desire that p may dispose us to undergo different mental states, and manifest in our stream of consciousness in different forms. It may give rise to an occurrent emotion of hope at the possibility that this desire might get satisfied, but it may also give rise to an episode of anxiety at the possibility that it will not get satisfied. If I have a desire, say, to preserve my physical integrity, to get home safe, etc., then in situations where I apprehend this desire as likely to get thwarted (for instance, by the presence of a menacing individual following me), then my desire to stay safe will manifest in the form of an occurrent episode of anxiety. I argue that this is exactly what happens to Suzy: it is because Suzy has a desire to preserve her physical integrity, that she is disposed to feel anxious when something puts her physical integrity under threat.

Note that all the reports that Stockdale defines as examples of “fearful hopes” express negative reactions to possible detrimental, unfavorable, menacing outcomes. As such, although they all manifest a desire (to be safe, etc.), thy express an awareness that this desire is at risk of being thwarted. Thus, while the term “hope” is used in all of these reports, the emotional experience that characterizes “fearful hopes” cases is more likely that of anxiety. Although the term “hope” is used to report them, what these individuals are reporting is their awareness of the possibility that they currently might be at risk of violence, loss, damage, etc. My understanding of these reports is that they all manifest an anxiety that one's desire to preserve one’s physical integrity might get thwarted.

“Fearful hopes”, I argue, are in fact anxieties. Stockdale herself, while describing Suzy’s mental state, writes:

Suzy will likely feel very anxious and panicky as she thinks to herself, ‘I really hope this man leaves me alone.’ But she will not feel pleasant feelings of anticipation about her hope’s obtaining” (2019, p. 121).

As I have shown, the fact that Suzy feels “anxious” and that she doesn't feel pleasant feelings of anticipation is not puzzling at all. Suzy feels this way because what she is in fact experiencing is primarily anxiety at the possibility of being assaulted. Note that the idea that Suzy is primarily apprehending her current situation with anxiety does not preclude her from feeling occasional glimmers of hope. It might indeed very well be the case that her state of anxiety at the possibility of being assaulted occasionally leaves way to a state of hope, with which it alternates. This is in fact an additional alternative explanation of Suzy’s state which is also theoretically less costly than Stockdale’s account, in that it does not require us to put the valence of hope into question. In other words, there are several ways to explain Stockdale's cases without having to commit to a view of hope as a negative emotion.

Taking stock, in the last two sections I have presented two different arguments for the thesis that hope and anxiety are not consistently negatively and positively valenced. According to the first account, anxiety is a mixed emotion due to its composed content: it consists in an alternation between our awareness of the possible good scenario, and the possible bad scenario. I have argued that this account of anxiety does not hold. Firstly, our psychological constraints do not allow for two incompatible hypothetical scenarios to get mixed (occur at the same time). Secondly, restless trepidation seems neither necessary nor sufficient for the experience of anxiety. I have suggested that the phenomenon of restless trepidation as described by these authors more plausibly corresponds to an alternation of anxiety with the emotion of hope in goal-relevant uncertain contexts.

According to the second account I have presented, certain occurrences of hope – “fearful hopes” – are negatively valenced as a result of occurring in menacing contexts and being partly constituted by fear. I have argued that, although we use the term “hope” to report them, these cases are not in fact manifestations of hope. This is because the use of the term “hope” in these reports expresses a desire (a disposition) and not an occurrent episode of hope. Moreover, emotional reactions to possible detrimental or menacing outcomes, which manifest the awareness that our desire to stay safe might get thwarted, correspond to episodes of anxiety, rather than hope. Desires to stay safe are precisely what dispose us to feel anxious about a possibility of threat to our physical integrity. I concluded that we have no reasons to think that certain occurrences of hope are negatively valenced.

Now, let us imagine that, at some moment on her way home, perhaps as she comes closer to her doorstep, Suzy feels a genuine episode of hope. She sees that her desire to get home safe has good chances of being satisfied, now that she is almost there. Suppose that what she feels is not relief at the idea that she is now out of danger, but genuine hope that she has good chances to reach her apartment safe and sound. If we part away this episode of hope from the other mental states going on in Suzy’s mind, how does this episode feel? Given that this hope occurs on the backdrop of a possible danger, isn’t the uncertainty constitutive of that hope experienced negatively? One additional way in which one could argue that this episode of hope is not entirely positively valenced is to say that, insofar as this episode of hope involves an evaluation of uncertainty, it is negative. In these kinds of cases, uncertainty itself is negatively evaluated by the individual. Certain circumstances make us averse to the mere awareness that we ignore some important aspect of our current situation.

As I will argue in the next section, this is again ignoring the fact that hope and anxiety are distinct precisely in virtue of constituting a positive evaluation or a negative evaluation of uncertainty. Whenever uncertainty is negatively evaluated, what we feel is anxiety. By contrast, hope results from a positive evaluation of uncertainty.

5 Polar Opposite Experiences of Uncertainty

As I have exposed in section II, the property best able to explain the valence of emotions is their evaluative nature (Carruthers, 2018; Teroni, 2018). Our emotions are positive and negative in virtue of the fact that emotions essentially constitute mental states which relate us to values, and that values are polarized entities.

Valence is thus an indispensable ingredient of what makes emotions the evaluative states that they are. The way in which we apprehend personally relevant values in our environment is through emotions, and it is in virtue of being valenced that emotions are these states which relate us to values. In other words, the valence of anxiety cannot be questioned on the basis of the presence of fleeting pleasurable affects interspersed within the experience. And the valence of hope cannot be questioned on the basis of the presence of fleeting unpleasurable affects interspersed within the experience. Valence is not a dispensable, contingent property of emotions: without their proprietary valence, these emotions would simply not be the intentional evaluative states that they are.

Now, what is peculiar about hope and anxiety is that they do not apprehend evaluative properties as instantiated in actual events: they apprehend evaluative properties that might get instantiated in possible events. They are reactions to goal-relevant possibilities. Nonetheless, hope and anxiety constitute polar opposite reactions to relevant possibilities. They present uncertain outcomes under radically different lights. Feeling hope is a way of engaging with the possibility of desirable outcomes an event may yield. Feeling anxiety is a way of engaging with the possibility of detrimental outcomes an event may yield. Hope and anxiety constitute two opposite stances towards uncertainty; two polar opposite ways of reacting to uncertain outcomes.

The fact that their intentional objects are uncertain raises specific issues – for instance, regarding the conditions of correctness of certain episodes of hope and anxiety (under which conditions is it appropriate, fitting, or warranted for us to hope or be anxious about a specific scenario?). However, what I want to clarify is that the uncertainty pertaining to the occurrence of their objects does not affect the valence of these emotions, because it does not bear on the fact that these emotions involve evaluative assessments of these possible scenarios. The crucial source of confusion in the debate on the valence of emotions of uncertainty seems to be the conflation between uncertainty of occurrence and uncertainty of value. In the case of hope and anxiety, although there is uncertainty concerning the occurrence of their intentional objects, there is no uncertainty concerning the nature of the evaluation that hope and anxiety provide of their objects.

Hope and anxiety constitute experiences of uncertain value in the sense that they are experiences of a value which might get instantiated, not in the sense that they are experiences of a vague, unclear, or unsettled evaluative property. Although the occurrence of the events that anxiety and hope take as objects are uncertain, the evaluation that anxiety and hope provide of their objects is quite clear. The possibilities that hope and anxiety take as their intentional objects are consistently assessed as either desirable or aversive, depending on whether we react to them with hope or with anxiety. While hope is an emotion which apprehends an event or state of affairs as uncertain and desirable according to the standard account of hope (Martin, 2016; Meirav, 2009; Milona, 2019), anxiety is an emotion which evaluates an event or state of affairs as uncertain and threatening (Kurth, 2018). In other words, if I desire to keep my bike, and I do not know whether my bike has been stolen, I will react to the possibility that it might have been stolen with anxiety, and I will react to the possibility that it might still be attached to the pole with hope.

The fact that hope and anxiety have uncertain states of affairs as their intentional object does not change the fact that these emotions constitute consistent assessments of these states of affairs as either favorable or unfavorable. While anxiety is the experience of an uncertain state of affairs as carrying a negative value for us, hope is the experience of an uncertain state of affairs as carrying a positive value for us. No matter whether these scenarios are actual or merely possible, hope and anxiety are specific ways of engaging with these scenarios; ways which present these scenarios as personally meaningful for us, and as instantiating either a positive, or a negative value. We can now understand why Stockdale’s suggestion (2019, p. 119) that: “when we hope that p and the content of p is perceived as threatening, we experience fearful hope” misses the point about what hope is. When an emotional episode evaluates a state of affairs as threatening, it is anxiety.

Hope and anxiety manifest a radically distinct evaluative orientation to one’s future. It is because hope is positively valenced that it puts us in an attitude of motivated pursuit towards the scenario of interest. And it is because anxiety is negatively valenced that it puts us in an attitude of motivated avoidance of the given scenario. In other words, the fundamental fact that hope and anxiety relate us to distinct values explains why they take the shapes that they do: hope takes the shape of pursuit, promotion, and longing because it is a way to engage with the possibility of desirable outcomes, and anxiety takes the shape of caution and vigilance because it is a way to engage with the possibility of threatening or detrimental outcomes. The respective valence of hope and anxiety is derived from the fact that these emotions constitute polar opposite ways of engaging with uncertain outcomes.

If this is valid, then there can be no such thing as ambivalent anxiety, or negative hope. When one feels bad about the threatening possibility of being assaulted on the street, this is anxiety; an anxiety which manifests the awareness that our desire to stay safe might get thwarted. And when one feels good about the possibility of winning the championship, this is hope – although one’s state of hope might alternate with a state of anxiety at the possibility that one in fact loses, giving rise to restless trepidation. The theory of valence presented explains why, and dissolves the confusion surrounding the valence of emotions of uncertainty.

6 Conclusion

In this article I have examined an idea present both in folk psychology and in certain philosophical accounts, according to which hope and anxiety are not strictly and consistently bivalent due to their epistemic profile. I have presented two philosophical accounts and shown that none of them is in fact tenable. A hope that is “infused” with or “tainted” by anxiety most plausibly expresses an episode of anxiety manifesting a background desire that no loss or harm occurs to one. Moreover, a state of “restless trepidation” most plausibly manifests an alternation of anxiety with hope, both likely to be elicited by high-stakes contexts with uncertain outcomes. Finally, the idea that hope and anxiety constitute polar opposite evaluative reactions to possibilities is key to explaining why they are respectively, and necessarily, positive and negative. Valence is the very way in which the (positive or negative) evaluation of their intentional objects is presented to us by emotions, and the fact that these emotions imply epistemic uncertainty about the occurrence of their target object should not tempt us to doubt their distinct and proprietary valence.