On the nature of implicit motives

Implicit motives as hidden states

As McClelland (1987) puts it, motivation has to do with the why of behavior (p. 4). As I explained in the introduction, the assumption seems to be that the answer to this question should be given in terms of a hidden state that causally explains the occurrence of the behavior. But of course, not all causes of behavior are motives: a complete answer to the question “Why?” also involves skill or trait variables, and cognitive variables (McClelland, 1987, pp. 4, 6). One important question, then, is how to distinguish motives from these other causes. To this aim, an (implicit) motive has been defined as a (nonconscious) concern, need, or disposition that drives, energizes, directs, and selects behavior toward the attainment of certain classes of goals, incentives, or outcomes (Brunstein, 2018, p. 369; McClelland, 1987, p. 183; McClelland et al., 1989, pp. 690–691; Schultheiss & Brunstein, 2010, p. ix; Schultheiss & Köllner, 2021; Winter, 1973, p. 17).

One problem with these definitions is that, as such, they are circular (see, e.g., Anscombe, 1963, pp. 10–11). A state or disposition is taken to be a motive because it brings about motivated behavior, and behavior is taken to be motivated because it is caused by a motive. But then we still do not know what either of them are. The same problem occurs when it comes to specific motives and their expressions. If it is the hidden cause that makes the action or thought an expression of the power motive, and it is at the same time the hidden state that is ascribed to the agent and receives its label “power motive” from the fact that the agent acts as such, we end up defining the motive in terms of the expressions it causes and the expressions in terms of what caused them. Either the motive or motivated behavior needs to be defined in a different way.

The question is how. In the psychological literature, the tendency is to elaborate on the nature of the motive. McClelland pays quite some attention to how we come to have certain motives (McClelland, 1987; McClelland et al., 1953). He argues that motives are built on primary, unlearned affect that is triggered by natural incentives that trigger affective arousal innately (McClelland, 1987, p. 110). On the basis of these innate combinations of incentives and affect, human beings adopt behaviors to sustain a positive state or to dissipate a negative one (McClelland, 1987). A motive, then, “is the learned result of pairing cues with affect or the conditions which produced affect” (McClelland et al., 1953, p. 75). But as Peters (1960, p. 44) has pointed out, this may be a correct explanation of how we come to have motives, but it does not tell us what, once they are learned, the nature of motives are. What is more, in order to take this to be an explanation of how we come to have motives, the phenomenon it explains has to be presupposed. Why is the result by definition a motive, and why can’t it be an explanation of how we acquire skills, for example?

Another strategy may be to connect motives to affect. McClelland et al. (1953, p. 30) state that affect is “obviously” important in controlling behavior, and Schultheiss (2008) defines the power motive as “the capacity to derive pleasure from having physical, mental, or emotional impact” (p. 606). But I do not think that motives themselves can be characterized in terms of affect. The idea is that the motive disposition, together with situational motive-relevant cues (Denzinger & Brandstätter, 2018, p. 3), causes a state of motivation. It is this state of motivation that is understood in terms of affect, and its occurrence is explained in terms of an underlying motive disposition (Schultheiss & Köllner, 2021). The motive, then, is not identified with affect, but is supposed to be its cause (McClelland, 1987, p. 128; McClelland et al., 1953, p. 35). But there are other causes of affect that we do not refer to as motives, for example physical contact with a loved one. Why should we think of these particular causes as motives? That they cause motivated behavior presumably, but we have already seen that is not going to work. A further problem is that affect does not distinguish between different types of motives. Whether we take an agent to be power motivated or affiliation motivated depends on the situation in which they experience pleasure or positive affect: is it when they are explaining things to others, or when they are working in a team? Affect is a sign of an agent being motivated, the motive itself cannot be identified with it.

Furthermore, even though affect may be a means to establish that an agent is motivated (McClelland et al., 1953, p. 30), why should we think that an agent is only motivated if they experience affect? McClelland et al. (1953) seem to think that “wanting” should be defined in terms of positive affect: “‘wanting to do a good job’ defines an end situation which would produce positive affect” (p. 39). But it seems to me that the fact that a boy is working hard and precisely is sufficient to conclude that he wants to do a good job; he does not need to feel good during the job or after it is finished as well (see Anscombe, 1963, pp. 67–68). If he is doing a good job, that is sufficient to conclude that he is personally involved (see McClelland et al., 1953, p. 79). One could of course argue that “real” motivation involves affect, but that cannot simply be assumed. A further problem with such a proposal would be that measures like the TAT do not even measure affect.

Another aspect that has been emphasized is that motives are aroused or triggered by cues or incentives (e.g., Brunstein, 2018, pp. 369, 377; McClelland, 1987, p. 137; Winter, 1973, p. 17). But I do not see how that could be a defining feature of motives. If the boy spontaneously decides to do a good job, we would not exclude his behavior as motivated. What is more, the trigger of the behavior is only labeled a cue or incentive and not a mere cause because it brings about motivated behavior and does not, say, cause a person to faint. Again, that means that the phenomenon that one is trying to explain, the motivated behavior, has to be presupposed. In order to take a cause to be a cue, we need to assume that its effect has to do with motivation.

This suggests that trying to make sense of motives as certain kinds of hidden internal states is not going to work. The attempts to pinpoint the unique character of motives without referring to the motivated behavior they are supposed to cause turn out to be unsuccessful. And if that is the case, we have also failed to capture the nature of motivation in general: it is unclear what the specific character of motivated behavior is and how it differs from other kinds of behavior.

Indeed, as Tanney (2009) explains when discussing Melden’s (1961) work, the problem is not only that the definitions are circular. The problem is that in ascribing a hidden state as an explanation of what the agent is doing, one has to already presuppose that what one is trying to explain is motivated behavior, or a state of motivation. And, in case of specific motives, one has to presuppose that the expressions in fact are expressions of, for example, the power motive. If an agent expresses themselves in line with what we take wanting to have an impact to be, if they dream about being famous or impressing others, for example (Winter, 1973, pp. 66, 69), the conclusion that this agent is power motivated is already drawn. The next step is to ascribe an unconscious state to this agent that causes all these expressions. But what does that exactly add? The hidden state is merely ascribed to the agent on the basis of these expressions. Even though, in principle, one psychological state may still causally explain the occurrence of the motivated behavior, it cannot be that in virtue of which the behavior is (power) motivated. The unique nature of motivation, motives, and motivated behavior can be found in the nature of the agent’s expressions. In interpreting these expressions—their actions, behaviors, remarks, emotions, thoughts, and their response on the PSE—we may come to the conclusion that the agent is motivated to have an impact. Ascribing a hidden state as a cause does not add anything to our understanding of motivated behavior in general or, more specifically, that the agent is motivated by power, for example. For that sake, it is redundant.

What is more, if we would insist that the existence of such a state is necessary to make sure that the agent acts for the power motive and not for a different motive, we would never in fact know whether that is the case. As Tanney (2009, p. 100) points out, even if we would try to measure such a state in a different way, for example via brain measuring techniques, we already have to have established that the brain activity, or other physical markers, indeed correlate with what we take expressions of the power motive to be. Again, the nature of the agent’s expressions has to be established first. On the basis of that, we decide whether the agent is motivated by power and whether we are justified in ascribing a hidden state to the agent. And if that is the case, ascribing a hidden state has no clear purpose anymore.

I can think of two reasons why someone would insist on ascribing such a state. First, one might assume that, in order to make sure that the agent is really motivated by power and not by something else, we need to know what caused the behavior (see McClelland, 1987, p. 22). The thought may be that we only truly understand a phenomenon if we know how it came about. But as I have argued, there is no reason to think that whether an agent is motivated by power is hidden from view. It is out in the open, right in front of our eyes, in the agent’s actual expressions. Ascribing a hidden state is not going to add anything in drawing this conclusion. Rather, it suggests that the view of implicit motives as hidden states is the result of unnecessary reification: “a tendency to take a characteristic of an ongoing process for the source of that process” (van Dijk, 2016, p. 994). A common characteristic of the expressions is identified, that they have to do with wanting power, and that aspect is turned into an object that is the cause of these expressions.

A second reason to think of implicit motives as causally efficacious hidden states may be that psychologists in the field are interested in explaining the differences between people. Why is it that Susie is playing the piano? And if she is doing that in order to impress others, why is she doing that? The thought seems to be that there must be something “inside” her that explains why power elements show up in what she does, says, and thinks. And indeed, the motive disposition or state is ascribed in order to explain the state of motivation that an agent is in, for example, in response to certain cues (see Denzinger & Brandstätter, 2018, p. 3; Schultheiss & Köllner, 2021).

I do not want to deny that we can empirically establish differences between agents that explain why one agent is motivated by power and another agent is not, or less. But we need to distinguish between different “Why?” questions. The answer to the question of why Susie is playing the piano can be given in terms of a motive, for example to impress others or to have power (Winter, 1973, p. 66). If we ask a further “Why?” question, that is, why she plays the piano in order to have impact, we seem to have two options. Either we explain it by elaborating on the fact that Susie is motivated by power, for example, “impressing others makes her feel good” (I say more about this option in the next section), or the answer is an explanation of why Susie is the kind of person that is motivated by power. In the latter case, we may refer to her motivational systems and associative networks (e.g., McClelland, 1987, p. 128; Schüler et al., 2015, p. 839) that are built up of past experiences (Schüler et al., 2015, pp. 840, 845). For example, we can claim that Susie is motivated by power because her parents permitted her to be aggressive when she was young (McClelland et al., 1989, p. 699). ² What is crucial, is that in this case we still provide explanations of what Susie does, but we do not explain it in terms of a motive. Rather, we give an explanation of why Susie has the power motive. Of course, research into how an agent comes to act for a certain motive is important and interesting, but we should make a distinction between what it is to act for a certain motive, and an explanation of why the agent acts for that motive. My claim is that the first cannot be captured by ascribing a hidden state, while the latter allows for all kinds of explanations, for example in terms of associations, past experiences, or physiological processes.

To conclude, in order to make sense of what implicit motives are, we should focus on the phenomenon on the basis of which an implicit motive is ascribed to the agent: the behavior toward the attainment of certain classes of goals, incentives, or outcomes.

Motives and the structure of action

The question now is: what characterizes goal-directed behavior? In virtue of what do we take behavior to be motivated or, more specifically, to be performed in light of a certain motive?

An account in which the nature of motivated behavior, or action, takes center stage and that is receiving a growing amount of attention, is Anscombe’s (1963; see, e.g., Ford et al., 2011; Schwenkler, 2019; Wiseman, 2017). Anscombe argues that actions have a means–end structure. When we act, we do one thing in order to do something else. Susie may be playing the piano in order to impress others, and the boy may be working hard and precisely in order to do a good job. Actions do not stand on their own; they are performed in light of an action at a higher level of description. What is more, agents have a distinct kind of knowledge of their actions: practical knowledge. Agents have this direct, nonobservational knowledge of their actions and their means–end structure, because their actions are caused by this knowledge: it is “the cause of what it understands” (Anscombe, 1963, p. 87). If the boy works hard and precisely with the intention of doing a good job, he knows that. And if Susie is playing the piano to intentionally impress others, she knows that too. If “doing a good job” and “impressing others” are their respective intentional actions at the highest level of description, that would be their direct answer to the question “Why?” They had the intention to do a good job or to impress others, and practically reasoned that working hard and precisely or playing the piano would be good ways to reach these respective aims (see, e.g., Ford, 2015, p. 138).

That does not imply, however, that agents cannot have ends they do not have practical knowledge of; it is perfectly possible that in acting or expressing themselves in other ways, agents have (further) ends that they do not know of. What is crucial is that if they lack practical knowledge of these ends, they do not intentionally act for them either.

According to Anscombe (1963, p. 70), the sequence of “Why” questions terminates when the answer to the question cites a so-called desirability characterization. At that point, it is clear why it is desirable for the agent to perform the action; the agent conceives of the end as being, in some sense, good. An end could be desirable because it is (a) useful: instrumentally good for the realization of some further end; (b) suitable: good in itself, given the agent’s vision of the good human life in general; or (c) pleasant: something the agent is aware of enjoying (Frey, 2019, p. 1139). That the agent conceives of the end as useful, suitable, or pleasurable explains why they are acting or motivated to act towards that end. That means that, in this view, in contrast to what is sometimes suggested in the psychological literature (e.g., McClelland et al., 1953, Chapter 2; Schultheiss, 2008, p. 606), pleasure and positive affect are only one way in which an action can be desirable for an agent. Both “doing a good job” and “impressing others” could be desirable in any of these three ways, and therefore they could count as desirability characterizations. If we would ask a further “Why” question, beyond the desirability characterization, we end up in ethics (“Impressing others is part of leading a good life”), or giving an explanation of why the agent cares about that end (“My parents taught me to always do my best”). The answer is no longer an answer in terms of the agent’s end in acting as such.

Anscombe (1963) offers a characterization of action that is not circular. Actions are caused by practical knowledge, and practical knowledge is insight into the means–end structure of that action. But how do the power, affiliation, and achievement motives fit into this picture? In Anscombe’s view, motives interpret actions (p. 19). Anscombe distinguishes between different kinds of motives (pp. 20–21), but for our purposes forward-looking motives are most relevant. Forward-looking motives can be described as the further intentions for which the agent is acting, that is, their end in what they are doing (p. 21). It can be captured with the phrase “in order to.” The boy is working hard and precisely in order to do a good job, and Susie is playing the piano in order to impress others. Impressing others is Susie’s motive for playing the piano.

At the same time, as we have seen, wanting to impress others is taken to be an expression of the power motive (e.g., Winter, 1973, p. 66). That suggests that the motives that McClelland and others distinguish are generic descriptions, for example, power, under which specific ends, for example, impressing others, an agent may act for fall. Describing Susie’s playing the piano in order to impress others in terms of the power motive gives us a more generic description of what she is doing. We interpret her action in a certain way, and that helps us make sense of why she acts as she does, mainly in connection to other actions, behaviors, and things she says, feels, and thinks that also reflect the power motive. That is why I use the phrase “may act for”: these more specific ends may also be ends the agent fantasizes about or that are reflected in their feelings and behaviors, which are also taken to be expressions of the power motive. The implication is that these motives are not causal explanations of actions, but are ways in which expressions (actions, feelings, thoughts, and behaviors) can be categorized and made sense of.

What, then, are implicit motives in this picture? Anscombe argues that agents have practical knowledge of their actions and ends. That may suggest that motives are by definition explicit. I think, however, that Anscombe’s theory of action paves the way for an alternative account of the implicit motives that are central in McClelland’s theory. We can see now how the power, achievement, and affiliation motives could be conceived of as implicit in a different sense, not as objects that are inaccessible through introspection. ³ An agent can act for a motive without being aware of doing so, if their action can be described in a certain way and they do not know that their action can be described as such. For example, if we know that Susie applied for a scholarship at the conservatory, her neighbor got it instead of her, and we see her slashing the tires of her neighbor’s car the next day, we can safely conclude that she slashed the tires out of jealousy. Acting out of jealousy is a pattern of conduct with which we are familiar (see Tanney, 2009, p. 103): we know what jealousy looks like. Whether Susie is aware that her action can be seen as an act of jealousy does not matter; Susie may have never heard of the concept of jealousy. Still, that was the motive with which she slashed the tires. Of course, Susie’s emotions and thoughts are part of the pattern too, but in such a clear case her personal report will not make much of a difference to our interpretation.

A similar account can be given of the implicit motives that McClelland and psychologists inspired by his work distinguish. Playing the piano in order to impress others is a “power act,” independently of whether Susie is aware that her action can be described as such. ⁴ In fact, Susie may not even be aware that she is trying to impress others. Susie could play the piano simply because she enjoys it, but Susie’s family may notice that she only enjoys it when lots of people are around, that the expression on her face resembles the look she has when she has won a game of chess, and that she always plays this one difficult song she can play very well. On the basis of that, they may (safely) infer that she enjoys playing the piano in this situation because it is a means to impress others. In such a case, she is trying to impress others and have an impact without being aware of it, without her knowing that she is aiming for this end.

In fact, when it comes to the motives that McClelland distinguishes, we have reason to think that they are implicit more or less by definition. Generally, people do not act under the description of “having an impact,” “being competent,” or “maintaining friendships.” Rather, they explain to their friends why to vote for a certain political party, they study hard, or they send out birthday cards to friends they haven’t seen for a while (see Williams, 1985, p. 10). “Having an impact” does not function as the end of an action, because it is too generic to play this role (Anscombe, 2005, p. 152). It cannot be the starting point of practical reasoning. For an agent to act, they have to have a more specific end; that is, to convince friends to vote for a certain party, or to participate in a talent contest. “Having an impact” is only a generic description of these more specific ends the agent does act for (or fantasizes, or thinks about).

Of course, Susie could at some point realize that the fact that she only plays the piano during parties or that she does it to impress others can be seen as an expression of the power motive. Similarly, the boy could realize that trying to do a good job reflects an achievement motive. But since this is not the end of their specific actions, they do not have practical knowledge of it. In this case, the agent, just like the psychologists in the field, needs to observe and interpret their action and other expressions to establish whether what they do, feel, or think counts as an expression of one of these motives. It requires that the agent interprets their own conduct. Thus, because of their general character, the power, achievement, and affiliation motive are for the largest part only implied in action, feelings, and thought, but not explicitly expressed. These implicit motives are not hidden from view because they are deeply buried in the agent’s mind, but because it is a matter of seeing an agent’s (or one’s own) conduct in a certain light that involves interpretation.

McClelland and other psychologists in the field are not mainly interested in the specific actions or expressions that I have been discussing so far. They want to say something about the agent, whether that agent has a certain motive. The proposal I have set out in this section may suggest that agents do not have motives in the sense that has been proposed in the psychological literature. Implicit motives are not personal states that can be found inside an agent and cause certain expressions. And if implicit motives are nothing more than generic descriptions of more specific ends, does it even make sense to think of an implicit motive as something that can be connected to the agent? I think it does, in the sense that many of an agent’s ends may reflect a similar motive (Meiland, 1963, p. 68). Similar to how we say of an agent that they are jealous if they often act out of jealousy, we can say that an agent is motivated by power if they often aim for (or think/fantasize about) ends that reflect a concern for having an impact. More generally, one could say that it reflects an agent’s “general conception of how to live, of what needs pursuing or avoiding in [their] life in a general way” (Frey, 2019, p. 1137). And indeed, wanting to have an impact, to be competent, or to maintain friendships are not just random descriptions, they do capture ends that many people seem to find or understand as desirable (McClelland et al., 1953, pp. 80–81). The more certain ends show up in the agent’s feelings, thoughts, actions, and behavior (and stories they write about pictures of social situations), then the more we would say that the person has a certain motive.

What is crucial is that in this view motives are not set in stone, as they would be on the hidden mental state picture. This is in line with McClelland’s (1987, p. 45) claim that one of the aims of scientific research is to establish whether categorizing ends as such will lead us to understand, explain, and predict human conduct better, or whether other categorizations would be more informative. This is also why I think we should avoid thinking of implicit motives in terms of dispositions (e.g., Ryle, 1949). Ascribing such dispositions to agents gives the impression that implicit motives have a definitive character. Furthermore, we would again ascribe a state to the agent that does not add to our understanding of implicit motives, or what it means to have a certain motive. My proposal would be that “having the power motive” is nothing over and above the agent displaying a certain pattern of what they do, feel, and think (see Kalis, 2019; Tanney, 2009).

Research on implicit motives

In the previous section I proposed to understand the power, affiliation, and achievement motives as generic descriptions of specific ends agents may act for. Because of the generic nature of these descriptions, these motives are more or less implicit by definition: they are only implied in the agent’s actions, feelings, and thoughts. Agents do not explicitly act for them. Whether an agent has a certain motive depends on whether the ends that fall under that description often show up in what they do, feel, and think.

I realize that this proposal may diverge strongly from how psychologists tend to think about motives. Because of that, even though I cannot provide a detailed analysis of how my proposal connects to the empirical findings, in this final section I point out that my view is in line with the most important empirical insights, and even explains some of them.

First, as I pointed out in the introduction, administering tests like the PSE involves interpreting the stories that participants write in response to pictures. On the view I defended, this is perfectly in line with the nature of implicit motives. Finding out whether an agent acts for or has a certain motive is a matter of interpreting their conduct, it does not involve finding out whether some hidden state is there in the mind of the agent. Also, it is clear why the motives measured in the PSE are implicit, even though the agent is aware of the stories they are writing and their writing may involve fantasies that they know they have. It is because the descriptions at the level of the motives are more generic than what an agent would explicitly write in response to one specific picture. The power motive is, perhaps with a few exceptions, only implied in what they are writing.

Furthermore, once we recognize that the power motive is not an internal cause but a way in which acts, behaviors, feelings, and thoughts can be interpreted, it should come as no surprise that correlations between different measures are low (Schüler et al., 2015). Since implicit motives are reflected in a pattern of expressions, it is likely that different tests and different pictures capture different parts of the pattern (Machery, 2016, p. 116). What is more, not all agents display the “full” pattern of the power motive. Some may often try to convince others but are never aggressive, while others fantasize about being famous but do not try to control what other people do. This is in line with Schultheiss and Pang’s (2007) suggestion, based on their experience in conducting experiments, to “use pictures that are similar to the situation in which you assess your dependent variables” (p. 331). These empirical insights reflect that the power motive is not one underlying cause, but is reflected in many different expressions that may often occur together (or not).

A criticism of my view may be that it does not capture the idea that implicit motives are affective needs (e.g., Brunstein, 2018, p. 369; McClelland et al., 1989, p. 691; Schultheiss & Pang, 2007, p. 322) that agents act for automatically (McClelland et al., 1989, p. 699; Schultheiss & Pang, 2007, p. 340), which may be blocked by social norms, lack of skill, and/or opportunity (Brunstein, 2018, p. 369). That is what measures like the PSE are supposed to capture. Indeed, in this paper I did not take the reasoning behind the PSE, that our motives influence the way we interpret the environment (Winter, 1973, pp. 33–34), as a starting point. However, my proposal is compatible with how measures like the PSE are taken to operate. The only difference is that it is not the implicit motive itself that influences the way in which the participants respond to the pictures. Other factors have to play this role, for example associative networks and memories (Schüler et al., 2015). Furthermore, I take it to be a strength of the view that the implicitness is connected to the expression itself, not to how it was caused. I do not think that the distinction between implicit and explicit can be captured by focusing on whether the expression came about automatically. For example, even if Susie decides to play the piano to impress the visitors without thinking beforehand, she may still be consciously playing the piano in light of that end. The fact that she acts automatically does not make her motive implicit, which implies that the way in which her action came about does not capture the difference between implicit and explicit motives (see Gawronski et al., 2006). What we should focus on when it comes to implicitness is how the agent relates to what they are actually doing, feeling, and thinking.

Furthermore, I think that the view of implicit motives as needs gives us even more reason to think of them as reflected in patterns. It has been argued that internal consistency is not a reliable indicator of reliability of PSE and similar measures, because a motivational process is “by its very nature not constant but characterized by a dynamic waxing and waning of need states, depending on environmental incentive cues and opportunities for need satisfaction” (Schultheiss & Pang, 2007, p. 324). That suggests that the presence of cues and opportunities are part of the pattern as well. A need is not some internal state; we ascribe it to the agent because of how they respond in certain situations, taking their history into account as well.

Finally, my proposal is also in line with the view that implicit motives can both be aroused—which has been used to establish the validity of the TAT (McClelland et al., 1953; Winter, 1973)—and at the same time reflect differences between people. All of us can have feelings and fantasies about having an impact, but some people have it more often and in more ambiguous situations. If anything, this suggests that there are several causal explanations of expressions of the power motive, and that interpreting them as reflecting the same motive is a matter of interpreting the expressions, not of finding one underlying cause.

An implication of my proposal is that measures like the PSE measure only the internal expressions of the pattern (see Alvarez, 2017), that is, what the agent fantasizes about but may not do or say. But in my view, external expressions—what the agent does say and do—are part of the pattern as well. To get a full picture, then, these elements of the pattern should also be taken into account, or at least cannot be ignored in assessing the validity of measures like the PSE. Do agents that score high on power have jobs in which they can make an impact? Are they frustrated or unsatisfied if they are told what to do? Are they generally concerned about their reputation? Do they tend to be angry or aggressive? These expressions are, from my standpoint, part of the same pattern; they also reflect whether the agent has a certain implicit motive.

Crucially, my proposal does not imply that the distinction between implicit and self-attributed motives should be rejected. The consequence is, however, that implicit measures like the PSE and self-report questionnaires do not measure two distinct mental states. Rather, the implicit measures capture (part of) the agent’s pattern of actual expressions, while the self-report questionnaires measure the agent’s own perspective on which ends they tend to or want to act for. Self-report questionnaires do not assess how an agent actually leads their life; they measure how the agent takes themselves to live their life. And, indeed, as Winter (1973, p. 32) points out, agents may not be very good at judging which ends in general they tend to or want to act for. Even though an agent has direct knowledge of their own actions, others are probably better at recognizing generic patterns in how they lead their lives and, with that, which implicit motive(s) they have.