Abstract

Abstract Counterfactuals are thoughts about alternatives to past events, that is, thoughts of what might have been. This article provides an updated account of the functional theory of counterfactual thinking, suggesting that such thoughts are best explained in terms of their role in behavior regulation and performance improvement. The article reviews a wide range of cognitive experiments indicating that counterfactual thoughts may influence behavior by either of two routes: a content-specific pathway (which involves specific informational effects on behavioral intentions, which then influence behavior) and a content-neutral pathway (which involves indirect effects via affect, mind-sets, or motivation). The functional theory is particularly useful in organizing recent findings regarding counterfactual thinking and mental health. The article concludes by considering the connections to other theoretical conceptions, especially recent advances in goal cognition. Keywords counterfactual thinking; regret; goals; rumination; mental simulation; inference; decision making; conditional; volition; motivation Who among us has never wondered about what might have been had some past choice been different? With perhaps a little more effort, you might have been an athlete, a doctor, maybe even a rock star. Who among us has never regretted choices made and actions taken? Maybe you should have studied harder in school, traveled more when you had the chance, or had the salmon for lunch instead of the pasta. And who has never pondered a lost love and imagined how passionate it might have been? Thinking about what might have been, about alternatives to our own pasts, is central to human thinking and emotion. Such thoughts are called counterfactual thoughts. Counterfactual thoughts are mental representations of alternatives to past events, actions, or states NIH Public Access Counterfactual thinking seems to be a common feature of people's conscious mental landscape THEORETICAL BACKGROUND The earliest theoretical tradition to explain counterfactual thinking was norm theory The functional perspective on counterfactual thinking, which is the focus of this article, emphasizes top-down than rather than bottom-up processes. In this view, counterfactual thinking may be seen primarily as a useful, beneficial, and utterly necessary component of behavior regulation. Accordingly, counterfactual thoughts are closely connected to goal cognitions, as we elaborate further throughout this article. Counterfactual thoughts are typically activated by a failed goal, and they specify what one might have been done to have achieved that goal THE FUNCTIONAL THEORY: OVERVIEW The two defining features of a functional interpretation of a psychological process are that (a) the process is activated by a particular deficit or need and (b) the process produces changes that end the deficit or fulfill the need. In the case of counterfactual thinking, if its primary function is problem solving, then counterfactual thinking should be activated by problems, and it should have the effect of evoking behaviors that correct those problems. This proposition is at root a regulatory loop-governing behavior. This regulatory loop is depicted in The process begins with a problem, mishap, or other negative experience that falls below a reference value for success or satisfactory performance. Recognition of a problem then activates counterfactual thinking (step 1 in Counterfactuals and Goals: An Illustration The connection of counterfactual to goals was illuminated in a simple demonstration conducted in our lab. The study was a modified 20-statement task in which undergraduate participants recorded 20 open-ended statements of "whatever comes easily to mind" in response to two kinds of stems (10 statements for each stem): a counterfactual stem ("if," with instructions to focus on "how the past might have been different") versus a causal attribution stem ("because of," with instructions to focus on "how certain events in the past brought about other kinds of events"). Thus, causal statements represented the control condition. After recording the 20 statements, participants were asked to mark a "P" or a "G" or both beside each statement to indicate whether it referred either to a personal action or a goal (defined to participants as "something intended; something someone works toward and wants to achieve" Mechanisms, Old School: Contrast and Causal Inference The regulatory loop described above was, in previous writings, connected to the second of the following two mechanisms underlying the consequences of counterfactual thinking: contrast effects and causal inference effects Contrast effects occur when a judgment becomes more extreme via the juxtaposition of some anchor or standard The second mechanism identified by Roese (1997) was that of causal inferences. Causal inference effects occur because a counterfactual conditional may emphasize, dramatize, or illuminate the causal link between an antecedent behavior and a desired outcome. To say that "If I had studied harder, I would have passed" is to underscore the causal impact of studying on grades. By virtue of their conditional structure and implicit reference to a parallel factual statement, counterfactual comparisons to actual sequences of events serve to isolate one particular causal antecedent in terms of its sufficiency to produce a divergent outcome. Although this partition of mechanisms of counterfactual thinking into either of contrast effects or causal inference effects served a previously useful explanatory role, recent evidence has rendered it somewhat less effective. First, affective assimilation effects, as well as contrast effects, have been demonstrated to follow from counterfactual thoughts Mechanisms, New School: Content-Specific Versus Content-Neutral Pathways We suggest a distinction between content-neutral and content-specific pathways from counterfactual thinking to action. Gollwitzer and Moskowitz (1996) used a similar distinction to describe the more general influence of goals on action, and given that counterfactual inferences are often goal related, this distinction provides a useful way to organize an expanding set of interlocking findings. The content-specific pathway involves the transfer of information (regarding action that might have been taken) from the counterfactual inference to behavioral intentions, which in turn influence performance of corresponding behavior. The content-specific pathway is what appears in THE CONTENT-SPECIFIC PATHWAY Step 1: From Problems to Counterfactual Thinking The regulatory sequence begins with the occurrence (and recognition) of a problem, mishap, or other negative experience that falls below a reference value for success or satisfactory performance. For example, a student might receive a failing grade on an exam, which represents a discrepancy between current performance and the reference standard of success (such as the goal of attaining a grade of B). Recognition of a problem then activates counterfactual thinking (step 1 in In highlighting problems as a key determinant of counterfactual thinking, we emphasize that this determinant influences the mere activation of counterfactual thinking, that is, whether an individual bothers to think "if only" versus merrily moving forward without a backward glance. A different set of determinants, such as norm violation and perceived control, has been shown to dictate the content of counterfactual thinking. These determinants have been reviewed elsewhere Step 2: From Counterfactuals to Intentions The content-specific pathway involves a particular insight contained in the counterfactual regarding the usefulness of some action, which is then transferred to a corresponding behavioral intention, which then fuels behavior. For example, if the counterfactual is "If only I studied more of the definitions for the anthropology exam, I would have performed better," then the content-related intention would be "I intend to study more of the definitions for the next anthropology exam." This direct semantic connection emerges because the specific causal insight contained in the counterfactual (an issue to which we turn below) provides the basis for the formation of a behavioral intention. The first experimental demonstration of the effect of counterfactuals on intentions was somewhat ambiguous, however. Roese (1994, Experiment 2) manipulated counterfactual thinking by having participants first focus on a recent academic performance that was disappointing, then generate three ways that the performance might have been better (upward counterfactual) versus might have been worse (downward counterfactual); an additional, no- NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript counterfactual, control condition was also included. The dependent measure was a set of Likert intention ratings. Participants who generated upward counterfactuals reported elevated intentions to perform future success-facilitating behaviors compared with no-counterfactual control participants. Generation of downward counterfactuals had no effect relative to the control participants. 1 Similar between-participant demonstrations of the effect of a counterfactual manipulation on behavioral intentions were presented in the contexts of computer purchasing Although the above studies may be taken as evidence of a content-specific pathway, ambiguity remains as to whether the effects involved formation of intentions inspired directly by the content of an upward counterfactual or resulted instead from a motivational push by the negative affect that sprung from that upward counterfactual (via a contrast effect). In the latter case, the motivational push would result in higher Likert ratings on a range of intention judgments, both specific and unrelated to the counterfactual in question. A study by Morris and Moore (2000, Study 1) is similarly ambiguous. These authors examined the near-accident reports filed with the Aviation Safety Reporting System by experienced pilots, then coded these narratives for both counterfactual content and "lessons learned," phrased as explicit statements by pilots regarding "an action that will be taken in the future" (p. 746) that would lessen the risk of another near accident. Upward, self-focused counterfactual statements were correlated with these intention statements, even after controlling for the severity of the original event. Intriguing as these studies were as demonstrations of the usefulness of counterfactual thinking for performance improvement, the ambiguity remains: Are intentions affected by counterfactuals by way of transfer of specific content? To address this ambiguity, a recent series of experiments provided more direct evidence for the content-specific pathway by which counterfactuals influence behavior. Smallman and Roese (2007) used a sequential priming paradigm to demonstrate that counterfactual thinking facilitates intentions to perform specific contentrelated acts. In each trial, participants made two judgments, the first involving a counterfactual versus control (or baseline) judgment and the second involving a behavioral intention judgment. A methodological challenge was how best to prime the "counterfactualness" of information while leaving constant its informational content. The solution in these studies was to manipulate only the stem but not the main body of an action phrase. In the prime segment of each trial, participants first saw an event, designed to establish the context (e.g., "spilled food on shirt"). Two seconds later, a stem plus action phrase appeared below, to which participants pressed a key to indicate their agreement or disagreement. In the counterfactual trials, a counterfactual stem was paired with the action phrase (e.g., "might have" + "eaten more carefully"). In the control trials, participants answered a question about the action phrase that was specified by the control stem. Control trials were intended to draw attention to the action phrase without encouraging inferences of a counterfactual nature. In one study, for example, participants answered questions about how many words appeared in the action phrase (e.g., "more than 1," "fewer than 4"). In another study, the stem prompted them to consider the frequency with which the action occurs in daily life, by way of a frequency stem paired with the action phrase (e.g., "common behavior?" + "eaten more carefully"). To achieve generality, several counterfactual stems were used (might have, could have, should have, etc.). Counterfactual versus control stems were interspersed randomly. The target task was a behavioral intention judgment semantically related to the negative event in the prime 1 The absence of a difference between the downward and the control condition might suggest that those in the control condition spontaneously generated downward counterfactuals. Evidence from other studies (e.g., Epstude and Roese Page 7 Pers Soc Psychol Rev. Author manuscript; available in PMC 2008 May 30. NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript task. On each trial, the stem "In the future I will" appeared at center screen; directly beneath it appeared an action that varied according to the content of the prime task (e.g., "eat more carefully"). Again, participants indicated their agreement with a key press. In addition, a noprime baseline condition was used, in which intention judgments were made without preceding prime judgments. Relative both to the control judgment and to baseline, counterfactual judgments produced a significant facilitation in response latency to respond to behavioral intention prompts. This effect was evident across several variations in procedure, including different control judgment tasks. These experiments revealed, not simply that a counterfactual might energize intentions by producing a heightened or stronger belief in a future course of action, but more specifically that the completion of a counterfactual judgment brings to mind information that facilitates the construction of a behavioral intention. Of course, one could argue that this effect might be motivational, in that for the brief moment in which counterfactual judgments were made, a temporary motivational impulsion energized intentions and would have had the same effect even on intentions not directly related to the counterfactual in question. A further experiment ruled out this possibility that a content-neutral mechanism could be sufficient to explain the facilitation effect. This experiment compared the facilitation effects on behavioral intentions that were matched versus mismatched in semantic content to the counterfactual prime. If a general motivational energization had produced the facilitation effect, that effect would be equivalent in both the match and mismatch cases. However, the facilitation effect was evident only when the counterfactual and intention matched in content, thus ruling out a motivational mechanism as the sole contributor to the facilitating effect of counterfactuals on behavioral intentions. Another experiment further clarified this effect by demonstrating that facilitation occurred on behavioral intention judgments but not on a different, control judgment (focusing on whether the action could be performed singly or required multiple actors), even though the two kinds of judgments were based on the same action content and involved response latencies of similar duration and variability. Overall, this research provided the strongest evidence to date that counterfactual thinking bears a close and special causal link to behavioral intentions. Step 3: From Behavioral Intentions to Behavior The next step in the content-specific pathway involves the link from behavioral intentions to behavior. Two kinds of evidence have appeared: indirect evidence demonstrating the effect of intentions on behavior, and direct evidence involving manipulations of counterfactual thinking on behavior as mediated by the content-specific pathway. Intentions influence behavior-In research inspired by the theory of reasoned action and the theory of planned behavior Counterfactuals influence behavior-Roese (1994, Experiment 3) used an experimental induction of counterfactual thinking (having participants generate and write down task-specific counterfactual thoughts on a piece of paper) positioned between the completion of two block of anagrams. This computer-presented anagram task was multifaceted, involving the acquisition of points through quick solution of each anagram, or rapid termination of the trial (when it appeared too hopeless to solve) so as to avoid penalty. In addition, participants could "buy" clues at a cost to the total score accumulated. The rules of this task were relatively straightforward and involved several decision points and possible strategies toward which counterfactual thoughts could be directed. Inducing participants to generate counterfactuals focusing on the first block of anagrams resulted in significant improvement on the second block of anagrams. Moreover, it was those counterfactuals that were upward (as opposed to downward) and additive (as opposed to subtractive) that facilitated performance. These performance benefits were linked to participants' identification, within their counterfactual statements, of performance-enhancing strategies (such as being quicker to skip past unsolvable anagram trials and buying fewer clues because their cost did not justify their value). Subsequent application of these strategies (recorded by the computer) was indeed associated with performance increases. Thus, when participants generated counterfactuals that focused on the addition of new actions that would have made their past performance better, the participants were more likely to act on them and to reap the ensuing rewards. Reichert and Slate Further evidence came from Morris and Moore (2000, Study 2a), who created an experiment in which participants attempted to land a virtual aircraft using Microsoft Flight Simulator software running on a desktop computer. A manipulation of counterfactual direction of comparison regarding previous landing attempts resulted in improved landings on subsequent trials. In addition to these experimental demonstrations, several correlational findings are consistent with the interpretation that counterfactuals influence performance. Landman, Vandewater, THE CONTENT-NEUTRAL PATHWAY The content-neutral pathway involves the way information is handled, as opposed to the particulars of the information itself. That is, independent of the meaning contained in a counterfactual thought, the counterfactual may also ignite attentional, cognitive, or motivational processes that themselves alter behavior. To clarify, in the content-specific pathway, apple-thoughts lead to apple-behaviors ("I should have eaten an apple" results in subsequent apple consumption), whereas in the contentneutral pathway, apple-thoughts might lead to orange-behavior (or turnip or squash or guava behavior). Three distinct of contentneutral effects-mind-sets, motivation, and self-inference-have been demonstrated.