Review of Philosophy and Psychology

, Volume 4, Issue 2, pp 293–308 | Cite as

Explaining the Effect of Morality on Intentionality of Lucky Actions: The Role of Underlying Questions



People’s moral judgments affect their judgments of intentionality for actions that succeeded by luck. This article aimed to explain that phenomenon by suggesting that people’s judgments of intentionality are driven by the underlying questions they have considered. We examined two types of questions: questions about why people act, and questions about how they succeed in acting. In a series of experiments, we found that people prefer different questions for neutral and immoral actions (Studies 1 and 2) and that asking them to think about questions they would not have preferred can change their judgments of intentionality (Study 3). These experiments suggest that neutral actions are judged to be less intentional simply because they do not motivate observers to ask questions which draw attention to the actor’s mental states. We discuss the potential application of this framework to other concepts affected by morality in surprising ways, such as causality.

“I wasn’t lucky. I deserved it.” – Margaret Thatcher

“Luck is what happens when preparation meets opportunity.” – Seneca

People might disagree about whether things that happen by luck are really attributable to the lucky person himself. Some, like Margaret Thatcher, might imply that luck is incompatible with merit; what happens by luck doesn’t happen by the actor’s choice. Others, like Seneca, might prefer to think that what we call “luck” is really the result of the hard work and choices of an actor. Which view of luck one accepts could lead to vastly different perceptions of lucky actions and the lucky people committing them.

This paper examines the effect of luck on how people perceive one feature of an action: intentionality of the actor. Do we really do anything intentionally when we manage to do it only by luck? As it turns out, people do not accept just one answer to the age-old quandary about the nature of luck when it comes to judging intentionality. They tend to be part-time Senecas, attributing lucky results to an actor acting intentionally, and part-time Thatchers, setting aside lucky results as flukes the actor carry out intentionally. Surprisingly, which philosophy seems right for any given lucky action turns out to depend on the moral valence of the action itself.

Over the last decade, researchers have confirmed the role of morality in people’s intuitions about the intentionality of a wide variety of actions. When the intentionality of an action is questionable for some reason (such as luck), the difference caused by the morality of the action becomes clear. Consider the stories used in the original Knobe (2003b) study that documented this effect. The neutral vignette describes a man participating in a sharp-shooting contest:

Jake desperately wants to win the rifle contest. He knows that he will only win the contest if he hits the bulls-eye. He raises the rifle, gets the bull’s-eye in the sights, and presses the trigger. But Jake isn’t very good at using his rifle. His hand slips on the barrel of the gun, and the shot goes wild… Nonetheless, the bullet lands directly on the bull’s-eye. Jake wins the contest (Knobe 2003b).

When participants are asked to read the story and decide whether Jake hit the bull’s-eye intentionally, only 28 % of people say yes. However, imagine the story is changed just slightly so that the action is terribly immoral:

Jake desperately wants to have more money. He knows that he will inherit a lot of money when his aunt dies. One day, he sees his aunt walking by the window. He raises his rifle, gets her in the sights, and presses the trigger. But Jake isn’t very good at using his rifle. His hand slips on the barrel of the gun, and the shot goes wild… Nonetheless, the bullet hits her directly in the heart. She dies instantly (Knobe 2003b).

Here, the man is equally awful at shooting and benefits from just as much luck, but the morality of the action has changed. In this case, did he hit her intentionally? 76 % of people say yes. Although the basic action and skill are the same, the morality of the action strikingly changes people’s intuitions about intentionality.

This type of effect has been widely replicated: morally extreme lucky actions (whether extremely moral or extremely immoral) are seen as more intentional than morally neutral ones. The effect occurs when the lucky action is typing in a random series of numbers (Nadelhoffer 2005) or rolling certain numbers on a die (Nadelhoffer 2004). It has been replicated using the shooting action even when the luck required to shoot successfully is extraordinary, with the shot bouncing off of various surfaces to hit its target (Sousa and Holbrook 2010). A similar effect of morality on attributions of intentionality has been even more thoroughly replicated, in which people judge the immoral side effects of actions to be more intentional than moral or morally neutral side effects; that effect has even been found cross-culturally (Knobe and Burra 2006) and in children (Leslie et al. 2006). It is clear that, although people might not expect morality to matter, it does. In this paper, I will present a new explanation for the role of morality involving the underlying questions observers consider as they assess the intentionality of the actor.

1 Existing Theories Explaining the Morality Effect

Why would morality play any role in judgments of intentionality? Researchers have suggested and rejected a number of different possibilities.

One set of hypotheses relates to the role of blame in judgments of intentionality. Some authors have posited that people’s judgments of intentionality are biased by their drive to justify blame for immoral actions (Alicke 2008; Nadelhoffer 2006). Another, slightly different view suggests ascriptions of intentionality stem from conversational pragmatics; people think that saying an action was not intentional implies that it was also not blameworthy, so they call an immoral action intentional even when they might honestly think it isn’t in order to communicate their assignment of blame (Adams and Steadman 2004).

Clearly the morality of the action somehow changes people’s perceptions of intentions, but is blame for immoral actions really what drives the differing ascriptions of intentionality? A number of different types of experiments suggest not. First, negative affect (as would be expected to accompany blame) is not necessary for people to exhibit this effect; morality plays a role in judgments of intentionality even among a population with brain damage impeding their ability to feel negative emotions (Young et al. 2006). Second, an experiment measuring reaction time for judgments of blame and judgments of intentionality found that people make decisions about intentionality faster than attributions of blame (although they decide about morality before either one); if blame were causing the intentionality judgments, it would have to be processed first (Guglielmo and Malle 2009). Third, the effect of morality holds even when all possible responses indicate the blameworthiness of the actor and only differ about his intentionality. Pellizzoni et al. (2010) told participants two judges had to decide on a punishment for someone who caused an immoral or moral side effect, and they had to choose which judge had the better argument; although both judges said that the actor was responsible for the harm and should pay a sanction, they disagreed about whether he produced it intentionally. In that situation, when the request to judge intentionality could not be interpreted as a question about blame, participants still found immoral actions more intentional. Finally, ascriptions of blame do not mediate the relationship between the morality of the action and judgments of intentionality (Sripada and Konrath 2011). These varying lines of research make clear that morality matters, but blame does not seem to be the correct mechanism through which it affects attributions of intentionality.

A few hypotheses remain, however, which do not rely on the desire to blame. One idea is that there are different possible meanings of “intentional,” and the moral valence of an action influences which definition people tend to use. Nichols and Ulatowski (2007) noticed a pattern in explanations of intentionality for immoral and side effects. Most people who find the side effect unintentional explain their judgments based on whether the actor had knowledge of the likely results of his action; however, most people who find it intentional explain it based on whether the actor wanted the results to occur. That pattern holds true even for the minority of people who find the immoral side effect unintentional or the neutral one intentional. Thus, even if everyone would agree in their assessments of both factors (knowledge and desire), prioritizing them differently for different moral valences drives the difference in ratings of intentionality. Cova et al. (forthcoming) extend this approach to the study of lucky actions, and similarly find that there are different possible meanings of intentionally, relying on either desire, autonomy, or skill. Knobe (2006) also previously accepted a similar view, arguing that people focus on different definitions (based on foresight, trying, or skill) for intentionality and blame, depending on the moral valence of the action.1

However, certain issues have been raised with this hypothesis. The primary problem is that it is unnecessarily complicated; when it is only applied to intentionality, it may seem plausible, but morality affects many other concepts as well. This hypothesis would have to involve two definitions of every word that morality affects, including “know,” “cause,” and “free will,” to meaningfully account for the vast effects of morality on people’s judgments. These parallels are even more problematic in that one of the two definitions proposed for intentionality – judgments of knowledge – itself depends on morality. It seems simpler and more plausible to search for a single explanation that underlies all of these otherwise surprising judgments, rather than explain each as an independent phenomenon, especially when their meanings are so linked.

However, Nichols and Ulatowski’s explanation could be more fundamentally boiled down to the idea that there are different questions people can ask about an action as they decide whether or not it is intentional. Some people ask, “Did he want the result to occur?” while others ask themselves, “Did he know the result would occur?.” Nichols and Ulatowski suggest that those different questions feed into different definitions of “intentionally,” but the same basic premise could be applied to another mechanism. For example, people might use a consistent definition of each word, but the differing questions they consider might lead to disparate judgments about whether that definition is met. Although it seems implausible that “intentionally,” “know,” “cause,” and all the other relevant words have multiple definitions depending on context, a simpler and more unified explanation drawing on Nichols and Ulatowski’s work could rely on the way people understand events to decide whether they meet consistent definitions.

A final possible explanation relies on the idea that morally extreme actions are more telling about the mind of the actor. When someone does something extremely moral or immoral, it seems to say more about his unique mind since it’s a surprising action other people might not have taken, but when someone does something neutral or normal, it seems like less of a reflection of underlying character and intentions. For example, imagine that you see someone wearing an academic gown at a graduation ceremony – you probably do not assume they have a unique penchant for academic robes. But imagine you see that same person dressed the same way in a coffee shop – you probably do assume they like to wear academic gowns (Uttich and Lombrozo 2010). Just as unusual actions seem to be more indicative of people’s preferences, they might also be seen as more indicative of the actor’s intentions. A related hypothesis suggests that immoral actions are more likely to be seen as reflections of a person’s “deep self,” the underlying aspects of a person that aren’t simply flitting preferences in the moment, and actions that reflect the deep self are seen as more intentional (Sripada 2010). Both these theories argue that when people act in unusual, norm-violating ways, those actions seem to more meaningfully reflect their underlying natures (including their desires, preferences, and intentions), and that connection to underlying, individual character increases judgments of intentionality.

Although this explanation could be correct, there is another possible interpretation of these results. Maybe it doesn’t matter how telling the observer actually finds the action, but the process of thinking about whether or not the action is telling drives higher judgments of intentionality. Even if people conclude the action actually is not very telling, the time spent thinking about the actor’s mental states makes the action seem more intentional. The more surprising the action, the more time an observer tends to spend considering whether it is telling and thus the more intentional it seems; however, it isn’t the moral valence or level of surprise itself that causes the effect. The explanation suggested in this paper builds on Uttich and Lombrozo’s idea that strange actions are more telling about the mind of the actor, but tests whether the process of trying to understand the mind of the actor could be the critical factor driving higher ratings of intentionality, even if upon reflection an action isn’t found to be very telling.

These seemingly disparate theories have something important in common: both Nichols and Ulatowski (2007) and Uttich and Lombrozo (2010) think that what matters is the conclusions people draw, such as whether or not a definition is met and whether or not an action is telling. But we could examine the same studies differently and consider the process of drawing the conclusions instead. What questions are asked to decide whether a definition is met, or whether an action is telling? This paper will focus on the questions people ask rather than the answers they conclude are correct, to show how the aspects of an action we question and consider – and not simply the answers we reach – affect how intentionally we find an action.

2 A New Explanation: Underlying Questions

In this paper, I will propose an alternative model based on the underlying questions people focus on as they analyze an action. Previous work in disparate fields, including psychology, linguistics, and philosophy, already suggests that underlying questions can be key to people’s opinions and behavior. In psychology, research about construal level has found that priming people with questions about either how or why a character would do something significantly affects how far into the future people perceive that action to occur (Liberman et al. 2007) and how politely they would address that character (Stephan et al. 2010). Similarly, in linguistics, research has shown how mental questions inform our interpretations of language. Imagine reading this sentence: “In St. Petersburg, officers always escorted ballerinas.” It could mean that in every situation in which a ballerina is escorted, it is by an officer; alternatively, it could mean that in every situation in which an officer escorts, he escorts a ballerina (von Fintel 1994). Which one seems correct depends on which question is being answered: “Whom do officers escort?” or “Who escorts ballerinas?” Finally, in philosophy, research has shown that underlying questions also affect what people think it means to “know” something (Schaffer 2007) and what explanations they find satisfactory (van Fraassen 1980). Across these diverse fields, a common thread emerges: when people have different opinions or perspectives, they might not simply believe in different answers to the same questions – they could also be answering different questions in the process of forming their opinions. The same notion about different questions could also be applied to understand conceptions of intentional action.

To apply this idea about how questions influence opinions to intentionality, consider two types of questions observers could have about an action. They could ask, “Why did the actor act the way he did?” or they might ask, “How did circumstances beyond the actor’s choice make the action successful?” Each of these questions can sometimes seem very relevant or very irrelevant to an action. For example, if you saw a news story about a politician winning an election, you would probably find the question of how circumstances played out to make him win very interesting, but you would not wonder why he wanted to win, since that seems obvious. On the other hand, if you saw a news story about a seemingly random murder, you would probably spend somewhat less time wondering how the killer carried out the brutality, and much more time wondering why he was motivated to do so.

The effect of morality on attributions of intentionality could plausibly be explained by this discrepancy in relevant questions, through two steps:
  1. (1)

    The morality of the action affects which questions people consider

  2. (2)

    Which questions people consider affect how intentional they judge the action to be.


If people rarely ask certain types of questions for neutral actions, and thinking about those questions elucidates intentionality, it would make sense that people tend to see neutral actions as less intentional.

The first step of this explanation relies on the morality of an action affecting which questions people consider. As the examples about news stories suggest, certain questions seem more or less relevant when the action at hand is particularly awful. Returning to the original vignette about the sharp-shooter, it’s easy to see the difference in intuitions. When the sharp-shooter shoots at a target, it seems irrelevant to ask, “Why did he want to hit it?” After all, wouldn’t anyone in his position have done the same thing? It seems much more interesting to ask, “How did he manage to succeed?” since it is rather surprising that he won, given his lack of skill. But when the sharp-shooter shoots at his aunt, a question like “Why did he want to shoot her?” suddenly seems very relevant; in addition to wondering how he succeeded, a reader also probably wonders what made him do such an awful thing.

The second step of the hypothesis shows how this difference in relevant questions causes differences in intentionality judgments: when observers focus on questions that draw attention to the actor’s mental states, they are more likely to be aware of the actor’s intentions and thus find the action intentional. For example, when the relevant question is, “How did the actor manage to succeed?,” the answer doesn’t invoke the actor’s intentions at all. He succeeded because he was lucky. The circumstances, not his intentions, answer the question. However, when the relevant question is “Why did he shoot at his aunt?,” it draws attention to the actor’s choice to act the way he did. That kind of question forces the observer to notice the importance of the actor’s decision (above and beyond his circumstances): the event hinged on the decision of the actor. That subconscious consideration of the actor’s intent, drawn out through consideration of certain questions, makes people judge the action to be intentional.

Thus, it seems plausible to suggest that people do consider different questions about immoral actions than neutral ones, focusing more heavily on the type of question that invites consideration of the actor’s mental states for immoral actions. The process of answering those questions draws attention to the actor’s intentions, leading to higher judgments of intentionality. This explanation aligns with our intuitions that morality does not seem like it should play a role in judgments of intentionality. Unlike some other explanations, this one does not involve morality’s changing the bar for what it means to be intentional: intentionality judgments always depend on whether or not the observer is aware of the actor’s intent; morality’s role is merely to increase or decrease the likelihood that the observer considers the types of questions which spark awareness of intentions.

The two steps of the hypothesis are tested in separate studies. Experiments 1 and 2 showed the first step – that morality affects which questions people find relevant – by demonstrating that people do in fact find questions which draw attention to the actor’s intentions more relevant for immoral actions than neutral ones. Experiment 3 showed the second step – that which questions people consider impact their judgments of intentionality – suggesting that considering questions which focus attention on the actor’s intentions increases judgments of intentionality, even for neutral actions.

3 Experiment 1: Readers’ Forced-Choice Preference Between Questions

The first study explored whether there is a difference in which questions people prefer depending on the morality of the action they read about. If people have different preferences depending on the morality of the action, that result would support the first step in the overall hypothesis: people think more about questions which highlight an actor’s mental states for lucky immoral actions, but they only consider questions that highlight the lucky circumstances for neutral actions. Using immoral and neutral vignettes about a sharp-shooter (from Knobe 2003b), this study tests whether people prefer questions about why the man shoots or how he succeeds in shooting; people were expected to prefer the “why” question in the immoral condition, since it draws attention to the actor’s intentions, causing the higher judgments of intentionality.

3.1 Method

Participants were recruited from Amazon’s Mechanical Turk website and paid 25 cents to take the survey. Each participant read a short vignette about a man named Jake successfully shooting something despite a lack of skill with a gun. There were two conditions which differed based on what Jake was shooting: a bull’s-eye (in the “neutral” condition) or his aunt (in the “immoral” condition). These vignettes were taken directly from Knobe’s original (2003b) study about the interaction between morality and skill and were listed in full in the introduction of this paper.

After reading the vignette, participants were asked which of two questions about the story they thought was “better.” The “why” question focused on why Jake took the action he did: “I noticed Jake aimed at the bull’s-eye [his aunt]. Why did that happen?” The “how” question focused on how Jake succeeded in the action he took: “I noticed Jake successfully hit the bull’s-eye [his aunt]. How did that happen?” As a similar measure, participants were also asked an open-ended response question about what else could have happened if Jake didn’t hit the bull’s-eye or his aunt. Responses were coded as to whether they referred to an alternative world where Jake didn’t shoot, or an alternative world where he shot but something else happened other than the bullet hitting its target. Examples of responses referring to a world where Jake didn’t shoot include “He could have got a job!” and “Ask her for money.” Examples of responses describing alternative worlds in which he still shot include “The gun could have backfired killing himself,” “He could have hit someone/something else,” and “He could have missed the target entirely.”

Then, participants were asked to rate their agreement with four statements on a seven-point scale from strongly agree to strongly disagree: “Jake shot the bull’s-eye [his aunt] intentionally,” “Jake thought carefully about whether to shoot the bull’s-eye [his aunt],” “Jake should be morally blamed for shooting the bull’s-eye [his aunt],” and “Jake seriously considered not shooting the bull’s-eye [his aunt].” As a check of their comprehension of the story, participants were also asked whether the bullet hit the bull’s-eye or aunt; in all stories, the correct answer was that the bullet did hit its target, and participants who answered incorrectly were excluded.

3.2 Results

81 participants (44 female, 37 male) completed the survey. 68 identified as Caucasian non-Hispanic, 8 as Caucasian Hispanic, 1 as African-American, and 4 as Asian or Pacific Islander. 43 were ages 18 to 35, and 38 were ages 36 and over. One participant was excluded for answering the comprehension question incorrectly.

Replicating previous experiments, participants in the immoral condition rated the action more intentional (M = 6.23, SD = 1.29) than participants in the neutral condition (M = 4.30, SD = 2.33), t(78) = 4.54, p < .0001. More importantly, participants more frequently preferred the “why” question in the immoral condition (47.5 %) than in the neutral condition (7.5 %), X2 (2, N = 80) = 12.66, p < .001, and more participants also described an alternative in which Jake didn’t shoot in the immoral condition (18 %) than in the neutral condition (0 %), X2 (2, N = 80) = 5.64, p = .02. Finally, there was a marginally significant difference between ratings among participants who picked the “why” question (M = 5.90, SD = 1.48) and those who picked the “how” question (M = 5.02, SD = 2.27), t(78) = 1.67, p = .10.

3.3 Discussion

Readers are clearly more aware of different aspects of the action when they read about immoral actions than neutral ones. Even in an unprompted free response question, the difference between the two conditions was striking: every single participant who read about the bull’s-eye contest wrote about an alternative world where Jake missed his shot, while 18 % of participants who read about the murder of the aunt were thinking of an alternative world where Jake didn’t shoot, and nearly half of participants in the immoral condition (45 %) chose the “why” question in the forced choice between questions. It seems that people might have implicit, potentially unconscious awareness of what features of an action are relevant; depending on the morality of the action, sometimes they are focused on the importance of the actor as the one who chose the course of action, and sometimes they are focused on the importance of other factors in allowing the action to succeed.

This study clearly showed that there is a difference in which questions people prefer depending on the morality of the action. However, it is not clear how relevant people find the questions in an absolute sense. Since the question forced participants to choose one question or the other, the results could be interpreted in a few ways. Participants in the immoral condition split nearly evenly between the two questions: do they find them both very relevant, or neither very relevant, or do some people prefer one strongly while others prefer the other? Participants in the neutral condition vastly preferred the “how” question: did they choose it because they find it extremely relevant, perhaps more so than the participants in the immoral condition, or did they choose it only because they found the “why” question so irrelevant comparatively? The next study will have participants rate the relevance of the questions without forcing them to choose between them, in order to better understand their preferences.

4 Experiment 2: Readers’ Preferences Between Questions Without Forced Choice

The previous study showed that people find different questions relevant in considering an action depending on its morality. It allowed comparison between the relevance of the two questions in each condition, but it did not allow comparison of the relevance of the questions between conditions by providing any sort of comparative scale. This study uses a non-forced-choice methodology to examine how relevant people find the questions about the same vignettes from Experiment 1, providing a comparison between conditions and permitting participants to express the perceived relevance of neither or both questions. Participants were expected to find both questions relevant to the immoral condition, but only the question about how the actor succeeded in the neutral condition.

4.1 Method

Participants were again recruited from Amazon Mechanical Turk and paid 25 cents to take the survey. They read the same vignettes as in Study 1, and were shown the same two possible questions, which they rated on a 7-point scale from “very irrelevant” to “very relevant.” They then rated their agreement with the same statements as in Study 1, including that Jake shot intentionally. The same comprehension check was also included, and participants who failed to answer it correctly were once again excluded.

4.2 Results

188 participants (80 female, 108 male) completed the survey. 152 identified as Caucasian non-Hispanic, 8 as Caucasian Hispanic, 11 as African-American, 13 as Asian or Pacific Islander, and 2 as Native American. 134 were ages 18 to 35, and 54 were ages 36 and over. 2 participants were excluded for answering the comprehension question incorrectly.

Intentionality ratings were higher in the immoral condition (M = 6.41, SD = 1.09) than in the neutral condition (M = 4.17, SD = 2.14), t(186) = 9.16, p < .0001. As shown in Fig. 1, relevance ratings for the “why” question were higher in the immoral condition (M = 5.56, SD = 1.66) than in the neutral condition (M = 4.22, SD = 2.06), t(186) = 4.92, p < .0001. Relevance ratings for the “how” question were higher in the neutral condition (M = 5.72, SD = 1.42) than in the immoral condition (M = 5.15, SD = 1.71), t(186) = 2.50, p < .02. There was a significant correlation between relevance ratings for the “why” question and intentionality ratings, r(186) = .21, p < .01, and a marginally significant correlation between relevance ratings for the “how” question and intentionality ratings, r(186) = −.12, p = .10.
Fig. 1

Relevance ratings by question and moral valence in study 2. Error bars show SE mean

4.3 Discussion

Morality was predicted to affect two related variables in this experiment: first, that people would rate immoral actions more intentional than neutral ones, and second, that people would rate questions about why an actor acted more relevant for immoral actions than neutral ones. The first result was unsurprisingly supported. As in nearly all previous research, people found immoral actions more intentional than neutral actions.

More interestingly, the second expected result was also clearly supported. This experiment clarifies the results of Study 1 to show that people find both “why” and “how” questions relevant to immoral actions, but only “how” questions relevant to neutral actions. As in Study 1, people who read about a neutral action found a question about how the actor acted more relevant than a question about why the actor acted. For immoral actions, people find both questions highly relevant, supporting the interpretation of the evenly split preferences in Study 1 as evidence of an even preference between the two (and not as the different preferences of two, evenly-sized groups).

5 Experiment 3: Manipulation of Readers’ Consideration of Questions

The previous two studies suggest that people naturally think about different questions when they analyze immoral and neutral actions. But the hypothesis of this paper would predict a more unusual effect as well: if people could be convinced to think about different questions, their ascriptions of intentionality would change. This study sought to provide a clearer test of the causative impact of thinking about a given question on judgments of intentionality by randomly assigning people to consider “why” or “how” questions to see if thinking about “why” questions increased judgments of intentionality even for neutral actions. The actual answers the participants provided to the questions are unimportant, but the act of answering them would ensure that participants had spent a bit of time considering them. The random assignment would overcome the small sample size in Study 1 and clarify whether thinking about certain questions has – beyond the merely correlational role demonstrated in Study 2 – an independent effect on intentionality.

5.1 Method

This study used a 3 × 2 between-subjects design, manipulating both the morality of the vignette and which question participants answered about the vignette. Participants were again recruited from Amazon Mechanical Turk and paid 25 cents for participating. All participants read one of the same two vignettes from the prior study, about a man shooting either his aunt or a bull’s-eye. Participants were then divided among three conditions based on the question they answered after the story: a “why condition” where they were asked, “Why did Jake aim at his aunt [the bull’s-eye]?”, a “how condition” where they were asked, “How did Jake hit his aunt [the bull’s-eye]?”, and a control condition with no question. Participants then rated their agreement with the statement “Jake shot the bull’s-eye (his aunt) intentionally.” Finally, participants answered a comprehension check question.

5.2 Results

239 participants (117 male, 122 female) completed the survey. 181 participants identified as Caucasian non-Hispanic, 15 as Caucasian Hispanic, 16 as African-American, 20 as Asian or Pacific Islander, 5 as other, and 2 declined to identify a racial group. 144 participants were ages 18 to 35, 71 were older than 35, and 24 declined to identify an age group. 11 of the participants were excluded for failing the comprehension check question.

The means for each condition are displayed in Fig. 2. A 3 × 2 analysis of variance on the intentionality ratings of the remaining 228 participants revealed a main effect for morality, F(1, 227) = 104.62, p < .0001, such that the immoral condition (M = 6.39, SD = 1.08) yielded higher intentionality ratings than the neutral condition (M = 4.41, SD = 2.21). There was also a main effect for question F(2, 227) = 3.08, p = .05, such that the why condition (M = 5.68, SD = 1.66) yielded higher intentionality ratings than the how condition (M = 5.10, SD = 2.19) and the control condition (M = 5.08, SD = 2.27). There was an interaction effect, F(2, 227) = 4.14, p = .02, such that the effect of question was greater in the neutral condition than the immoral condition.
Fig. 2

Intentionality ratings by question and moral valence in study 3. Error bars show SE mean

Planned t-tests between the different neutral groups revealed greater intentionality ratings in the why condition (M = 4.97, SD = 1.95) than the how condition (M = 3.63, SD = 2.13), t(74) = 2.84, p < .01. There was also a significant difference between the intentionality ratings in the neutral why condition and control condition, t(69) = 2.27, p = .02. There was no significant difference between the neutral how and control conditions, or between any of the immoral conditions (all p-values > .05).

5.3 Discussion

This experiment shows the predicted result about the effect of questions: participants who answered a question about why Jake acted the way he did found his action more intentional than participants who thought about how he succeeded in his action. Since participants were randomly assigned to a question condition, this suggests that the effect is not merely a correlation but a causal connection. It seems that briefly drawing awareness (through a question) to an aspect of a story which otherwise would have gone unnoticed makes people more aware of it, which changes their intuitions about intentionality.

These data provide an initial response to the implication in Uttich and Lombrozo’s work that more telling actions are seen as more intentional (Uttich and Lombrozo 2010). Instead, this experiment suggests it is the process of thinking about the action and deciding whether it is telling which increases judgments of intentionality. The answers to the question about why Jake shot the bull’s-eye rarely mention anything unique about him; typical responses included “He wanted to hit it to win the contest,” “So he could win the rifle contest,” and “Because he needed to hit it to win.” These responses frequently used words that imply he didn’t have much choice, such as “must,” “had to,” and “only option.” Yet respondents still find his action more intentional than those in the control group simply because they have considered the question, even though their answers have little to do with his unique character or deep self. Apparently, it is the asking of the question, not the answer, which drives the effect.

In the neutral conditions, the effect of question was driven by the difference between the why condition and the other two conditions, with the how and control conditions producing roughly equal judgments of intentionality, as predicted. In Studies 1 and 2, when participants were allowed to choose which question they preferred, most participants chose the how question, suggesting that people tend to be aware of that aspect of the story even if they are not prompted to consider it. Thus, it makes sense that participants who were left to consider the story without guidance also focused more on aspects drawn out by the how question than the why question, and so their responses were similar to participants who were told to think about those same aspects.

It should also be unsurprising that there wasn’t an effect for the immoral condition. Studies 1 and 2 taken together show that people automatically consider both “how” and “why” questions when they think about immoral actions, but they find only “how” questions relevant for neutral actions, even if they are not forced to pick just one relevant question. Focusing on the “why” question may be critical to judgments of intentionality, but people were thinking about the “why” question in all three conditions; unlike for neutral actions, people don’t need to be prompted to be aware of that aspect of an immoral action. Given that people already considered the “why” question without prompting, it makes sense that active encouragement to think about it would not make much of a difference.

6 General Discussion

Studies 1 and 2 found that people prefer different types of questions depending on the morality of the lucky action they are thinking about: for immoral actions, they consider questions about why the actor acted and questions about how the action succeeded, but for neutral actions, they consider only the latter. Study 3 found that thinking about “why” questions raises judgments of intentionality, even for neutral actions. Taken together, the three studies provide support for the overall hypothesis that these underlying questions explain the effect of morality on attributions of intentionality. Unprompted, people tend to focus solely on how actors accomplish morally neutral lucky actions: their answer tends to focus on the unusual sheer luck, crowding out consideration of questions involving the actor’s decisions and mental states. However, immoral actions draw attention to questions that highlight the actor’s decisions in addition to the lucky circumstance, since norm-violating choices are just as surprising as good luck. That focus on questions which highlight the actor’s importance to the outcome increases judgments of intentionality; when people are encouraged to ask that type of question about neutral actions, they find neutral actions more intentional too.

Taken together, the studies provide a fairly intuitive explanation of the morality effect. As noted at the very start of this paper, it seems counter-intuitive that morality would matter to judgments of intentionality, and people do not tend to expect that it would. Studies on the morality effect appear to show precisely that morality matters to intentionality attributions, but this question-based approach provides a more intuitive alternative. The morality of the action doesn’t change people’s definition of intentionality – actions are always judged to be intentional if the observer is thinking about the intentions of the actor. All the morality does is change the likelihood of that thinking happening. More morally extreme actions are more likely to provoke that kind of thought, but as long as the observer considers questions whose answers highlight the actor’s mental states, ratings of intentionality can be high for neutral actions too. Thus, the question-based approach explains the empirical results about morality and intentionality without violating the widespread intuition that morality doesn’t seem like it should matter.

This same approach could potentially explain other phenomenon about intentionality and causality as well. In the vignette examined in this paper, the potentially relevant questions are “Why did the actor act?” and “How did the action succeed?,” but other situations could be explained by the varying relevance of a different question that draws attention to the actor’s mental states. For example, this approach could explain the classic side-effect effect of the chairman authorizing policies in spite of their moral or immoral side effects (Knobe 2003a). In that case (unlike in the lucky action case), most people are probably considering the actor’s mental states in some sense, but the morality of the side effect might make them ask about different mental states. When the chairman signs off on an evil side effect, they might think, “Did he know his policy would cause such harm?” but when his policy causes something good, they ask, “Did he want that good thing to happen?.” People might generally agree – regardless of the morality of the action – that the chairman knew his policy would cause the effect and that he didn’t particularly want it to. The difference between the conditions isn’t the answers to the questions but rather the asking of different questions.

This explanation would also account for the otherwise surprising result in a variation of the chairman case where the chairman has been misinformed about the effects of his actions. In that case, the Knobe effect disappears: people judge that a chairman who caused a bad side effect thinking it would be good did not cause it intentionally (Pellizzoni et al. 2010). The explanation based on underlying questions could account for such a result in two possible ways. First, perhaps respondents consider whether the chairman knew he would cause harm, as they normally would when they encounter an immoral action; since the answer is clearly that he did not know, they found the action unintentional. Second, perhaps people do not even consider the knowledge question much, because the vignette focuses so strongly on the chairman’s lack of knowledge. Instead, as they normally do for positive side effects, they consider the desire question and reach the same conclusion they would reach there: the chairman didn’t want the side effect to happen, so he didn’t cause it intentionally. Either of these changes in people’s internal questioning behavior could explain why the chairman’s ignorance undercuts the Knobe effect.

As another example, this approach could explain the phenomenon of “causal deviance,” in which people judge actions to be less intentional when they happen in different ways than the actor intended (Mele and Cushman 2007). Past research found that people judge a woman to kill weeds much more intentionally if she drives to the store to buy weed-killer and sprays them with it than if she is on her way to the store to buy weed-killer and accidentally drives off the side of the driveway, killing the weeds with her car. Why would it matter how exactly she killed the weeds if she had the same intentions in both cases? The underlying question approach provides a potential explanation: when an intention is carried out in a strange way, it shifts attention to questions that have nothing to do with the actor’s mental states (for example, “How did her car slip off the driveway to kill the weeds?”) rather than questions about the actor’s thoughts (for example, “Why did she want to kill the weeds?”), so people do not think about the intentions of the actor as much, and consequently rate the action less intentional. Similarly, this approach could explain the effect of morality on judgments of causality, in which people think that immoral actions cause a consequence and neutral ones do not, even when both actions contributed to the consequence (Knobe and Fraser 2008). For example, if a receptionist offers free pens to administrative assistants but not to professors, and she ultimately runs out of pens after both an administrative assistant and a professor take one, people judge the professor’s immoral pen-taking to have caused the shortage of pens much more than the administrative assistant’s permissible pen-taking. The questions approach could provide an explanation: when the professor does something he is not supposed to do, observers find the question of “Why did he take a pen?” quite relevant, but when an administrative assistant takes a pen as she is supposed to, observers don’t wonder why. Answering that question draws attention to the role of the actor’s choices in causing the ultimate problem of the pen shortage; when people disproportionately consider that question for the professor’s action, they then rate the professor to have caused the action more. Just as with intentionality, morality doesn’t necessarily change the definition of causality. It merely changes the likelihood that people think about the questions that cause them to think the definition has been met.

Beyond the issue of morality and intentionality, the idea of looking at what questions people find relevant could have a broader application. Researchers often look at what people think the answers to questions are, but in many fields, they have not yet considered what questions people are asking unprompted. However, the questions people ask in the real world might matter to explain the way they act. For example, perhaps people act in prejudiced ways when they don’t consider questions about a person as an individual, but rather as a member of a disadvantaged group. It might be that their answers to questions about the individual could be unprejudiced, but those questions never get asked in their head because they are focused on giving prejudiced answers to questions about the group. As another example, consider the bystander effect, in which crowds of people do not help when they see strangers in need (Latane and Darley 1970). It seems unlikely that everyone, whether in a small or large group, considers, “Should I save that innocent victim?” and the people in the large group happily conclude that they shouldn’t. Perhaps instead the people in the crowd refrain from helping because they don’t ever ask themselves the question of whether or not they should. They don’t consider the issue and choose not to help; they just never consider it because they assume someone else will. In these types of circumstances, understanding people’s answers alone isn’t as telling as knowing which questions get asked.


  1. 1.

    Knobe (personal correspondence) now rejects this view in light of the data provided in the present manuscript.


  1. Adams, F., and A. Steadman. 2004. Intentional action in ordinary language: core concept or pragmatic understanding? Analysis 64(282): 173–181.CrossRefGoogle Scholar
  2. Alicke, M. 2008. Blaming badly. Journal of Cognition and Culture 8(1): 179–186.CrossRefGoogle Scholar
  3. Cova, F., Dupoux, E., and Jacob, P. forthcoming. On doing things intentionally. Mind and Language.Google Scholar
  4. Guglielmo, S., and Malle, B. F. 2009. The timing of blame and intentionality: Testing the moral bias hypothesis. Unpublished manuscript, Brown University.Google Scholar
  5. Knobe, J. 2003a. Intentional action and side effects in ordinary language. Analysis 63(3): 190–194.CrossRefGoogle Scholar
  6. Knobe, J. 2003b. Intentional action in folk psychology: an experimental investigation. Philosophical Psychology 16(2): 309–324.CrossRefGoogle Scholar
  7. Knobe, J. 2006. The concept of intentional action: a case study in the uses of folk psychology. Philosophical Studies 130: 203–231.CrossRefGoogle Scholar
  8. Knobe, J., and A. Burra. 2006. The folk concepts of intention and intentional action: a cross-cultural study. Journal of Cognition and Culture 6(1–2): 113–132.CrossRefGoogle Scholar
  9. Knobe, J., and B. Fraser. 2008. Causal judgment and moral judgment: Two experiments. In Moral psychology vol. 2: The cognitive science of morality: Intuition and diversity, ed. W. Sinnott-Armstrong, 441–448. Cambridge: MIT Press.Google Scholar
  10. Latane, B., and J. Darley. 1970. The unresponsive bystander: Why doesn’t he help? New York City: Appleton-Century Crofts.Google Scholar
  11. Leslie, A.M., J. Knobe, and A. Cohen. 2006. Acting intentionally and the side-effect effect. Psychological Science 17(5): 421–427.CrossRefGoogle Scholar
  12. Liberman, N., Y. Trope, S. McCrea, and S. Sherman. 2007. The effect of level of construal on the temporal distance of activity enactment. Journal of Experimental Social Psychology 43(1): 143–149.Google Scholar
  13. Mele, A.R., and F. Cushman. 2007. Intentional action, folk judgments, and stories: sorting things out. Midwest Studies In Philosophy 31(1): 184–201.CrossRefGoogle Scholar
  14. Nadelhoffer, T. 2004. The butler problem revisited. Analysis 64(3): 277–284.CrossRefGoogle Scholar
  15. Nadelhoffer, T. 2005. Skill, luck, control, and intentional action. Philosophical Psychology 18(3): 341–352.CrossRefGoogle Scholar
  16. Nadelhoffer, T. 2006. Bad acts, blameworthy agents, and intentional actions: some problems for juror impartiality. Philosophical Explorations 9(2): 203–219.CrossRefGoogle Scholar
  17. Nichols, S., and J. Ulatowski. 2007. Intuitions and individual differences: the knobe effect revisited. Mind and Language 22(4): 346–365.CrossRefGoogle Scholar
  18. Pellizzoni, S., V. Girotto, and L. Surian. 2010. Beliefs and moral valence affect intentionality attributions: the case of side effects. Review of Philosophy and Psychology 1: 201–209.CrossRefGoogle Scholar
  19. Schaffer, J. 2007. Knowing the answer. Philosophy and Phenomenological Research 75(2): 383–403.CrossRefGoogle Scholar
  20. Sousa, P., and C. Holbrook. 2010. Folk concepts of intentional action in the contexts of amoral and immoral luck. Review of Philosophy and Psychology 1(3): 351–370.CrossRefGoogle Scholar
  21. Sripada, C. 2010. The deep self model and asymmetries in folk judgments about intentional action. Philosophical Studies 151(2): 159–176.CrossRefGoogle Scholar
  22. Sripada, C.S., and S. Konrath. 2011. Telling more than we can know about intentional action. Mind & Language 26(3): 353–380.CrossRefGoogle Scholar
  23. Stephan, E., N. Liberman, and Y. Trope. 2010. Politeness and psychological distance: a construal level perspective. Journal of Personality and Social Psychology 98(2): 268–280.CrossRefGoogle Scholar
  24. Uttich, K., and T. Lombrozo. 2010. Norms inform mental state ascriptions: a rational explanation for the side-effect effect. Cognition 116(1): 87–100.CrossRefGoogle Scholar
  25. Van Fraassen, B. 1980. The scientific image. Oxford: Oxford University Press.CrossRefGoogle Scholar
  26. Von Fintel, K. 1994. Restrictions on quantifier domains (Doctoral dissertation). University of Massachusetts, Amherst.Google Scholar
  27. Young, L., F. Cushman, R. Adolphs, D. Tranel, and M. Hauser. 2006. Does emotion mediate the effect of an action’s moral status on its intentional status?: neuropsychological evidence. Journal of Culture and Cognition 6(1): 291–304.Google Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2013

Authors and Affiliations

  1. 1.Department of PsychologyYale UniversityNew HavenUSA
  2. 2.Stanford University, Law SchoolStanfordUSA

Personalised recommendations