Skip to main content
Volume 14


Eva-Maria Konrad

Goethe-University of Frankfurt

Thomas Petraschka

University of Regensburg

Christiana Werner

University of Göttingen

1        Introduction

In this paper, we want to examine the claim that emotional responses are relevant for the understanding of literary texts. This claim can be encountered in different versions. We will distinguish between a weak and a strong variant, both of them prominently presented by Jenefer Robinson. When Robinson suggests “that our emotions help us in the construction of a satisfactory summary reading of the novel or play or movie, or what is often called an ‘interpretation’”[1], she advocates a weak version of the claim, because emotions are considered to be merely a non-necessary means to facilitate understanding.

We have no objections against this claim. There are numerous rather uncontroversial ways in which emotions can ‘help’ us to understand literary texts. Let us suppose John read Tolstoy’s ‘War and Peace’, and that it made a deep impression on him. Subsequently, John was curious about other books written by Tolstoy and read ‘Anna Karenina’ as a result. One could say that epistemic emotions like curiosity, interest, the urge to broaden one’s horizon, etc. played a role in understanding the work, because without them, John probably never would have read—and therefore, if all went well, understood— ‘Anna Karenina’. Therefore, it seems perfectly reasonable to us to say that emotional indeed helped John understand ‘Anna Karenina’.

But Robinson is not satisfied with the claim’s profoundness and proceeds to advocate a stronger version:

Even if you agree with me that an emotional experience […] helps us to understand [a novel], you might still think that it’s possible to come to the same understanding by a more cerebral engagement with the text. […] I’ll try to show that this is not true: nothing else can do the job that emotions do. Without appropriate emotional responses, some novels simply cannot be adequately understood.[2]

The last sentence of the passage contains what we consider to be the most ambitious and also the most interesting variant of the claim, stating that emotions are relevant for an understanding of literary texts. It is a strong variant, because it considers emotions to be a necessary condition of understanding.[3] In the following, we will concentrate on this version of the claim and analyse three central arguments Robinson presents in its favour.

2        The ‘Trigger’ Argument

The first argument could be called the trigger argument.[4] Its basic idea is this: every reader of a literary text is confronted with an enormous quantity of information she has to process. Emotions, so the argument goes, provide orientation and guidance in negotiating this thicket: They are “sources of salience”[5] in the sense that they “alert us to important aspects of the story”.[6] Therefore, if we react emotionally to a certain episode, event, or character, “we are in a good position to try to discover why we respond emotionally as we do, and this in turn can lead us to seek in the work the origins of this response. […] [T]hereby [we] acquire a deeper, fuller understanding of the work”.[7]

The trigger argument has some persuasive power indeed. Aspects that cause strong emotional reactions are very often important for an understanding of the text. Consider, for example, Anna’s last encounter with her son Seryozha in Tolstoy’s ‘Anna Karenina’. Not only does this episode usually elicit pity, the observation of Anna and her awareness of what her son means to her also causes the reader to realise that, from this moment on, Anna’s life will be nothing but miserable since she simply cannot live without her son. Additionally, the trigger argument explains why repeated reading of the very same text can potentially reveal new information again and again. Every time we read, we are in a different emotional state and our emotions bring different aspects of the text into focus, thus potentially providing new insights.[8]

Nevertheless, the potential trigger-function of emotions does not prove the strong claim that when it comes to an adequate understanding of literary texts “nothing else can do the job that emotions do”.[9] To support this claim, the trigger argument has to provide evidence that being triggered by her emotional reactions is the only way for a reader to become aware of significant episodes of a text. We cannot see how the trigger argument can do this. Think, for example, of a reader who is perfectly trained in all areas of literary studies. Her awareness could be triggered by a number of textual aspects that do not solely rely on emotional reactions, but are remarkable with respect to form or content as well: significant episodes might, for example, be placed right in the middle of the book (for example the fatal encounter of the English and the Scottish queen in Schiller’s ‘Mary Stuart’); they might be first or last instances of their kind (like the last encounter between Anna and Seryozha); or there might be special narratological features such as a change of focalisation or a change of discourse time (as for example the alternation of present and simple past at the beginning of Thomas Mann’s ‘The Magic Mountain’). To take one’s emotional responses into account is hence to draw conclusions from one of many possible sources of salience. Emotional reactions are not exclusive in this respect.

Additionally, it seems problematic to suppose that, as Robinson says, only “appropriate emotional responses”[10] can trigger our awareness of significant passages of a text.[11] Suppose a sadist reads ‘Anna Karenina’ and because of her sadistic constitution she does not feel pity for Anna but enjoys her sadness and pain instead. Why should the inappropriate emotion of joy trigger the reader’s awareness of central episodes of the text (like Anna’s last encounter with Seryozha) any less than the emotion of pity would? Apparently, any emotion can function as a trigger. The trigger argument therefore fails to prove the strong thesis that “[w]ithout appropriate emotional responses, some novels simply cannot be adequately understood”.[12]

3        The ‘Empathy’ Argument

Robinson considers a second way in which emotional reactions are necessary for understanding works of literature, namely emotional engagement with fictional characters:

The argument in a nutshell is this: (1) Understanding character is essential to understanding great realistic novels I have in mind; (2) understanding character is relevantly like understanding real people; and (3) understanding real people is impossible without emotional engagement with them and their predicaments.[13]

In the following we will take (1) for granted but discuss claims (2) and (3). In order to examine the assumed epistemic value of emotional engagement with a real person we will start with Robinson’s claim (3).

The phrase ‘emotional engagement’ refers to a wide range of emotional reactions, like falling in love with someone or being angry at one another; some reactions include forms of perspective-taking or feeling for or with another person, such as sympathy, emotional contagion, and empathy. Given the wealth of emotional reactions and the brevity of this article it is necessary to narrow down the scope of this phrase. Since it is often claimed that empathy is a way of understanding another’s mind or at least a process that helps us to understand others,[14] empathic emotions seem to be the most promising candidates for emotional reactions that might play a decisive role in understanding a character or person.[15]

Following Coplan, we understand empathy as an imaginative process in which an empathiser simulates a target’s mental situation exactly, where the latter involves not only states like beliefs and desires but also affective states.[16] Picking up Robinson’s proposal, one could hypothesise that empathy’s affective components, i.e. empathic emotions, have an epistemic value. We can, according to this specification, reformulate (3): Understanding real people is impossible without empathic emotions (towards them).

How can empathic emotions be a source of understanding others? Understanding someone’s mental situation involves information about her mental states, which in turn involves not only cognitive but also affective states. The phenomenology of an affective state is a central aspect of these mental states. If one does not know what it is like to be, e.g. in the state of grief, one lacks important information about grief and hence about the mental situation of a grieving person. Experiencing an affective state is a way of getting information about what it is like to be in that particular state. We do not want to argue here that it is metaphysically necessary to experience a certain affective state in order to know everything about the phenomenology of this state, but we claim that it is psychologically necessary: in order to acquire all the information about the phenomenology of an affective state, human beings have to experience this state.

Even if this is true, we still have not shown that experiencing a particular affective state empathically is necessary in order to get information about what it is like to be in that state and hence for a proper understanding of the other’s mind. Think of the following situation: A tries to understand B while B is in the affective state c, and A has experienced c before. For a proper understanding of B, is it necessary for A to experience c again, this time empathically? We do not think so: If A has had the experience of c before, she knows what it is like to be in a state of c (let’s take it for granted that she can remember the experience). If this is true then an empathetic experience of an affective state does not provide additional information about the phenomenology of the affective state in question if the empathiser has prior experience of the respective affective state. As a result we claim that empathic emotions are only necessary for those empathisers who have not felt the affective state in question before in order to fully understand the other’s mind.

Up to this point we have only focused on the epistemic value of empathic emotions in understanding real people. Now, we see that even if (2) is true, readers do not necessarily have to empathise with the fictional characters in order to know what it is like to be in their mental situation—a prior experience of the affective state in question is sufficient.

Let us have a look at premise (2) now. The question is whether understanding a character’s mental situation is relevantly similar to understanding the psychology of a real person. Even though there is no consensus on the question of which mental process the term “empathy” refers to, it seems that all accounts presuppose that there is a target with mental states. Independent of whether we follow a realist or antirealist account regarding the ontology of fictional characters, the problems for the fictional case start with this basic assumption: according to antirealist positions, there are no such things as fictional characters. Hence, there are no such things as mental states of fictional characters, whereas according to realist-creationist accounts, fictional characters exist. However here fictional characters are seen, e.g., as abstract artefacts[17] or posits of literary criticism.[18] In any case they are supposed to be entities without mental states. If one of these accounts is correct, either fictional characters, including their mental states, do not exist, or they only exist as entities without mental states. Against this background, the target of empathy in the fictional case is an open question. If we allow non-existent targets of empathy we still need to clarify what the mental states in question are.

A solution to this problem might be that empathisers in the fictional scenario simulate or imagine the mental states the fictional characters are in according to the text, e.g. if the fictional target is in a state of fear the empathiser has to simulate this particular state. Hence, the reader gets information about the phenomenology of the character’s particular state. But many literary works are not or at least not always explicit about the mental states of the characters. Therefore, figuring out the mental states of the fictional character turns out to be a matter of textual interpretation. Being directly confronted with a real person whose mental state we want to figure out, however, is not very different from this procedure. It is, of course, not a matter of textual interpretation, but still, we need to interpret her outer appearance and/or her behaviour. It is worth mentioning that even real targets of empathy are sometimes not perceived directly. An empathiser can empathise with a real person while or after reading a text about her. Thus, it seems that interpretation is involved in all these cases. So, we assume that in this respect there are no decisive differences between the real and the fictional case.

It is, however, important to point out that if there were two or more equally plausible interpretations of the outer appearance and/or behaviour of a real person, the interpretation would not have an influence on the fact that a real person is in a particular state (or that she is not). This seems to be rather different in the case of fictional characters. There might be fictional texts with more than one equally plausible interpretation with regard to the mental situation of a character at a certain point in the story. Moreover, there could be an interpretation a, according to which the character is in affective state x but not in affective state y. According to interpretation b, however, the character is not in affective state x but in affective state y. Even though the interpretations are contradictory, they might both be equally adequate, or equally “optimal”, as Gregory Currie puts it.[19] Readers will then, following these different interpretations, simulate different mental situations and will therefore get information about different mental states by means of empathy with a fictional character. One empathising reader, following interpretation a, will get information about what it is like to be in state x; another empathising reader, following interpretation b, will get information about what it is like to be in state y. As long as empathy is based on equally adequate interpretations in both cases, either piece of information the readers get as a result of the empathic process is equally epistemically relevant concerning the mental situation of the fictional character,[20] while in the case of a real target there is a fact of the matter that decides the epistemic relevance of the information.

As a result, we claim that empathy with fictional characters is decisively different to empathy with real people in relevant aspects. We have argued, nevertheless, that empathy with real people as well as with fictional characters can provide information that is necessary for a proper understanding of the fictional character or real person in question. However, empathy is not a unique way of gaining the sort of information discussed above. A prior experience of the affective state in question can be a source of this information as well. Hence, in order to understand (certain) literary texts properly, empathy is only necessary for those readers who lack prior experience of the affective state in question.

4        The ‘Data’ Argument

The results of our discussion of the argument from empathy are furthermore relevant for the third argument we shall now discuss. Its core idea is this: to adequately understand a literary text, it is not enough to grasp the content of the story and to process all the information given by the text in a purely cognitive way. According to the argument, there are additional features of literary texts that can only be grasped via emotional reactions:

[I]n reading a complex novel such as Anna Karenina, we don’t just emote about Anna. We use our emotional responses towards her as data in arriving at an interpretation of her character.[21]

As discussed above, empathic emotions toward fictional characters can be sources of information for those readers who have never experienced the affect in question before. However, Robinson seems to be thinking of a further kind of information that is relevant for an interpretation of a character and hence of the whole work, but is not necessarily related to a characters’s mental situation. In the following, we will focus on the question of which additional kind of data our non-empathic emotional responses might provide.

In the course of interpreting a text we often ascribe properties to texts or to fictional characters. Robinson ascribes the properties of being amusing and being repulsive to Strether and Macbeth, the protagonists of Henry James’ novel ‘The Ambassadors’ and Shakespeare’s play ‘Macbeth’, respectively.[22] Properties like these can be referred to as response-dependent properties, because whether an entity has such a property depends on a corresponding response. Roughly speaking, because we are repulsed by Macbeth’s ruthless striving for power and amused by Strether’s awkward yet loveable behaviour, Macbeth is repulsive and Strether is amusing. If the response which decides the issue is an emotional response—just as in the cases of being amusing or repulsive—it seems very plausible to say that our emotional reactions provide access to a special kind of information, viz. information on whether a fictional character has got a certain response-dependent property or not.

We are not, however, convinced that this amounts to proof for the strong claim. It is not clear to us why information about characters having certain response-dependent properties should be provided only by emotional reactions. We think it is possible to find out whether a character has a certain emotional-response-dependent property or not without reacting emotionally to the character oneself. For the sake of simplicity, let us suppose that a character has the property of being amusing or repulsive as soon as at least one person finds him amusing or repulsive. It is important to note, however, that although being amusing or repulsive depends on the emotional reaction of at least one person, it does not depend on a specific person’s emotional reaction. For this reason, we think that a reader can also find out that Macbeth is a repulsive character or that Strether is an amusing character if she, for example, observes the behaviour or the facial expression of another reader. Suppose Mary smiles, chuckles, and maybe sometimes even laughs out loud when she reads this or that passage of ‘The Ambassadors’ (and also suppose that Mary’s facial reaction and behaviour is connected to her emotional reactions in a reliable way), such that the reader can infer that Mary finds Strether amusing, and consequently, that Strether has the response-dependent property of being amusing. The reader’s own emotional reactions to Strether do not play any role in this scenario. If this is correct, information about emotional-response-dependent properties can, at least in principle, not only be gained by means of emotional reactions but also by other means.

To sum up: Although Robinson forcefully argues for the claim that emotional responses are necessary for an adequate understanding of literary texts, we remain doubtful about the strong version of her claim. While emotional reactions do of course often trigger our awareness of important episodes in literary texts, the reader’s attention can in principle also be grasped by other peculiarities of the same episode, which are perceivable without reacting emotionally. Robinson is also right in arguing that empathic reactions to fictional characters can grant information about the phenomenology of the affective states they are in. Empathic emotions, however, are not the only source of this particular information; prior experiences of corresponding emotions are sources as well. Therefore, empathic reactions are only necessary to fully understand the fictional character in question for those readers who have never experienced the emotion in question before. Finally, we have tried to show how it is possible to gain information about characters who have certain response-dependent properties even if one does not exhibit an emotional reaction oneself. If what we say is correct, the claim that emotional responses are necessary for an adequate understanding of literary texts is in need of additional support.



About the Authors    Eva-Maria Konrad is a Research Assistant at the Department of Literary Theory and Comparative Literature at the Goethe-University of Frankfurt, Germany. Her research interests concern fiction, fictionality and knowledge, and the philosophy of literature. She is currently working on counterfactual literature.

Thomas Petraschka is an assistant professor of German literature at the University of Regensburg, Germany. His research focuses on the philosophy of literature and German literature from the 19th and 20th century. Currently he is working on the relationship between literature and emotions.

Christiana Werner is leader of the junior research group “Language, Cognition and Text” at the Graduate School of Humanities Göttingen at the University of Göttingen. Her main research focuses on the philosophy of mind and philosophy of the emotions in particular and the philosophy of fiction. Currently she is working on an epistemological project about the affective components of interpersonal understanding.


Coplan, Amy. (2003). “Understanding Empathy”. In: Empathy: Philosophical and Psychological Perspectives, Amy Coplan, Peter Goldie (Eds.). Oxford: Oxford University Press, pp. 3–18.

Currie, Gregory. (2003). “Interpretation in Art”. In: The Oxford Handbook of Aesthetics, Jerrold Levinson (Ed.). Oxford: Oxford University Press, pp. 291–306.

Elgin, Catherine Z. (2008). “Emotions and Understanding”. In: Epistemology and Emotions, Georg Brun, Ulvi Doğuoğlu, Dominique Kuenzle (Eds.). Aldershot: Ashgate, pp. 33–50.

Miall, David S. (1989). “Beyond the Schema Given. Affective Comprehension of Literary Narratives”. Cognition and Emotion 3(1), pp. 55–78.

Novitz, David. (1987). Knowledge, Fiction & Imagination. Philadelphia: Temple University Press.

Robinson, Jenefer. (2005). Deeper than Reason: Emotion and its Role in Literature, Music, and Art. Oxford: Oxford University Press.

Robinson, Jenefer. (2010). “Emotion and the Understanding of Narrative”. In: Companion to the Philosophy of Literature, Garry L. Hagberg, Walter Jost (Eds.). West Sussex: Blackwell Publishing, pp. 71–93.

Stueber, Karsten R. (2006). Rediscovering Empathy. Agency, Folk Psychology and the Human Sciences. Cambridge, MA: MIT Press.

Thomasson, Amy. (1999). Fiction and Metaphysics. Cambridge, UK: Cambridge University Press.

Van Inwagen, Peter. (1977). “Creatures of Fiction”. American Philosophical Quarterly 14(4), pp. 299–308.

[1] Robinson 2005, p. 105.

[2] Ibid., p. 107. Robinson restricts her claim to “realistic novels, plays and films […] as part of the ‘Great Tradition’ of the English novel” (ibid., p. 106). In order to avoid an undue generalisation of Robinson’s thesis, we follow Robinson in her choice of examples and concentrate on realistic novels and plays as well. Additionally, it should be noted that it is not the aim of this paper to discuss general conditions of adequate understanding. However, we assume, like Robinson, that an understanding of literary works can be more or less adequate. Accordingly, once again we will follow Robinson’s own suggestions of what an adequate understanding might be in the examples she considers.

[3] Such a strong claim is not only advocated by Robinson, but also supported for example in Novitz 1987, p. 74f., and Miall 1989, p. 65. We suppose that understanding and appreciation are to be distinguished and that a full understanding of a literary text is possible without appreciating (parts of) the work.

[4] Cf. Robinson 2005, p. 108ff. & 126ff. Cf. also Elgin 2008.

[5] Elgin 2008, p. 43.

[6] Robinson 2005, p. 107.

[7] Ibid., p. 111.

[8] This is emphasised in Elgin 2008, p. 45.

[9] Robinson 2005, p. 107.

[10] Ibid. (our italics).

[11] What appropriate emotional reactions are exactly and what makes them appropriate is a difficult matter that we need not delve into deeply here. It seems quite clear that some notion of “appropriate emotional responses” is firmly embedded in our everyday behaviour (cf. Elgin 2008, p. 38). It is appropriate to feel joy after finally reaching a goal one has worked hard to attain and to feel pity when we see someone suffering—but surely not the other way round. Since the literary examples we take into consideration are mostly clear-cut cases like this, we do not intend to dwell on complicated borderline cases. Robinson herself also goes for this modest approach and explicitly restricts herself to “examples in which it is pretty clear that the emotional responses are appropriate to the text and were probably intended by the author” (Robinson 2005, p. 108).

[12] Robinson 2005, p. 107. As Robinson admits, the trigger-function of emotions alone cannot grant an adequate understanding of a literary text either: By triggering the reader’s awareness, emotional reactions apparently only help the reader to know where she should start her interpretation, i.e. which aspects, scenes, or episodes of a text she could take into consideration; but they do not help her with regard to the content of this interpretation itself, i.e. to interpret the emotionally triggered data in a reasonable way. A reader whose awareness is perfectly triggered by emotions could still be completely misguided in her interpretation of a text. To result in an adequate understanding of a literary text, the trigger must be followed by a thorough process of interpretation (cf. ibid., p. 122: “It is important to distinguish among experiencing the work, reflecting on our experiences of it as they occur, and interpreting it by reflecting on and reporting our experiences of the work after finishing it, by summing it up as a whole.”).

[13] Robinson 2010, p. 78.

[14] Cf. Stueber 2006.

[15] Other emotional reactions will be discussed in the following section on the ‘data’ argument.

[16] Cf. Coplan 2011.

[17] Cf. Thomasson 1999.

[18] Cf. van Inwagen 1977.

[19] Cf. Currie 2003, esp. p. 293.

[20] It is in fact an open question whether both readers empathise with one or rather two different fictional characters.

[21] Robinson 2005, p. 116.

[22] Cf. ibid., p. 111ff.