The Agentic Value of Intrusive Thoughts 

3.1. Social Sustenance in Agency

While some see agency as an individual endeavor, whereby one internally reflects and deliberates about one’s values, makes an autonomous decision, and proceeds to act on this decision, others have called these individualist assumptions into question. Instead, some have put forth an alternative approach to agency that is facilitated by social relations and collaborative interactions. Under this approach, one’s social environment plays an important role in shaping, sustaining and facilitating the expression of one’s values.

While it’s been suggested that other people can impact the formation of our values from the get-go (Doris and Nichols 2012; Alfano 2013; Roberts and Krueger 2021), agency is also intertwined with social relations by the sustaining and expressing of our values. Others often provide us with social or moral feedback, reinforcing certain norms and expectations, placing us into certain social roles. Stohr (2019) calls these people our “moral neighbors”—those whose presence and affirmation of certain norms (and disapproval of others) help hold us into certain roles or identities, including the aspirational self, or the self we wish to become. Our moral neighbors help support us in acting out our aspirational moral values when we might otherwise falter. A colleague calls you out when your gossiping is out of line; your HIIT instructor encourages you to stick to your exercise goals; following a lecture, your student asks if you are vegan. Within close relationships, psychologists have studied what they call the Michaelangelo Phenomenon, whereby one partner helps to ‘sculpt’ another, using positive reinforcement to ‘chisel away,’ better enabling one to enact their ideal self (Drigotas et al. 1999; Rusbult et al. 2009; Kumashiro et al. 2006). In short, others around us can provide us with support and reinforcement, aiding in the sustaining and expressing of our values.

Those who impact our agency vary from strangers to acquaintances to close others. People we’ve never met, but serve as exemplars, also help in the sustaining and expressing of our values. These exemplars might even be deceased individuals or fictional characters. Philosophical literature on the role of moral exemplars, and the experience of admiration and emulation in response to them, can be seen as emphasizing the ways that agency is socially supported (Zagzebski 2017). Experiencing this admiration comes with the motivation prompting us to take steps towards being more like this person. Exemplars, and more generally, those around us, can provide us with an image of what manifesting our values might look like. And, experiencing admiration can create a motivational force, prompting us to act on and express these values.⁶

But exemplars are not always positive—those around us show us not just what a morally good life would look like, but also a “not so good” (Stohr 2019, 88) life as well. Waggoner (2021) has further developed this point, arguing that negative exemplars can help us exercise our values, and so agency. Making use of the notion of one’s feared self, or “the selves we are afraid of becoming” (Markus and Nurius 1986, 954), Waggoner highlights the ways that others can provide us with an image of who we don’t want to be and instill in us strong motivations of fear towards acting in immoral ways.⁷ One’s feared self—as well as one’s ideal self—are two kinds of possible selves, or selves that one either fears or wishes to become (Markus and Nurius 1986). Just as other people can provide us with an image of our ideal or aspirational self, giving rise to admiration which motivates us to seek that which we value, others can also serve as examples of our feared selves. Whether these are actual others we have relationships with, strangers, or even imagined or fictional characters, negative exemplars can also show us what we do not value, or who we do not want to become. This recognition, likewise, comes with a motivational component—namely that of fear, or feeling repulsed by the values that another embodies and the actions they take. In this way, agency is sustained and expressed via social means: others provide positive and negative images of what we could be, thereby motivating us to stick to and express our values.⁸

3.2. UITs Help Us Uphold Our Agency

In this subsection, I first highlight that the content of UITs often reflects our values, and that experiencing UITs can function to help us act on our values, specifically by caring for those we love. UITs can provide a mechanism which reinforces our values, acting as an ‘other’ and serving the same sort of role that others typically provide.

First, UITs typically involve content that is relevant to our goals, cares, concerns, and loves (Moulding et al. 2014). If one has a UIT that involves the content of drowning their baby, this is usually an indication that one really cares for and loves one’s baby. Freeston and Ladouceur (1999) observe as much: “It is no coincidence that we typically see harming [UITs] among gentle people…[and] distressing sexual thoughts among highly moral people” (370). The more someone cares about taking or avoiding a certain action, the more prone they are to experience UITs with the content that depicts those cares or goals being thwarted. Thus, UITs are not meaningless, random, or disconnected from who one is and what one loves. Rather, the specific content of the UIT points to what one loves. This perhaps is one reason why UITs are experienced as so disturbing—“they scare because we care” (Moulding et al. 2014, 280).

Second, given a social approach to agency, UITs can play a similar role as other people, functioning as an ‘other’ to help one act on their values, and so uphold their agency. In particular, UITs serve as a negative exemplar, providing us with an image of our feared self. Interestingly, research shows that those who have a conception of, and are motivated by, their feared selves, are also more likely to experience UITs (Melli et al. 2016; Llorens-Aguilar et al. 2022). This makes intuitive sense, since UITs often depict one doing something very bad, or imagining oneself as acting in a way that one greatly fears could become a reality. Having a feared self as part of one’s self-conception can be particularly helpful in goal achievement when one’s fear motivates the person to stay vigilant, creating plans and taking steps away from one’s feared self, ensuring that it does not become a reality (Fuglestad et al. 2008; Fuglestad et al. 2013). Consider the report by Chasson et al. (2024) of the experiences of new mothers in the postpartum period, who had UITs about their infants; note the mothers’ emotion of fear:

• “It’s the fear of holding something so small in the water. Not to drop him, not to hurt him … not to get water in his mouth, not to drown him” (119)

• “When I’m standing near a barrier where there’s a sharp drop on the other side, maybe an upper story of the building near the stairs, I’m afraid that by accident I’ll throw the baby over it” (ibid).

Experiencing this fear can play an important role in motivating one to stay vigilant in situations where one’s UITs depict harm coming to the infant—such as simply holding one’s infant a little tighter while walking down the stairs in response to the fear of dropping them.⁹ Fairbrother et al. (2022) looked at UITs that occur postpartum in new parents, and suggested that they play an important role in helping a new parent better care for their infant:

[UITs] cues new parents to behave in ways to protect their infant from danger, with the ultimate goal of ensuring infant survival…UITs of infant-related harm (even UITs of intentional harm) likely trigger protective, safety-seeking behavioral responses and consequently increase infant safety…From what we know about new mothers’ behavioral responses to UITs of both accidental and intentional harm, it seems clear that this is the case; both types of UITs of harm are associated with behaviors intended to increase infant safety (102).

Given this important function that UITs can serve, UITs could play the role of another ‘person’ helping to uphold our agency. How might this work? First, one’s UITs might provide an image of their feared self, serving as a possible self to react against. This emotionally laden reaction can motivate one to increase their commitment about how they will act. Second, this commitment can and often does manifest itself in additional concrete vigilant behaviors. As a new parent now facing the possibility that one could end up harming their child, UITs are a way to guard against this possibility becoming a reality.

Thus, UITs sometimes express our values, but in a rather indirect way. Rather than a UIT about drowning one’s child manifesting a lack of love, a thought like this might actually express one’s love for their child, prompting them to take proper caution and care in protecting their child from danger. In the next section, I will go on to explain this agentic function or value of UITs is connected to their prevalence: It’s not just that UITs manifest our feared selves, but that they often do so precisely in the situations where bringing our feared self to mind is of great value for sustaining our agential capacities.¹⁰ In one’s normal environment, where one is surrounded by others, the presence of other people provides one with peer pressure, influence, checks, implicitly communicated norms, and feedback, holding one to one’s values. But when these constraints are absent, UITs also tend to be more prevalent. UITs are more apt to arise in situations of social isolation and uncertainty, which are also the kinds of situations we’d expect agency to otherwise be impaired. Thus, in these kinds of circumstances, UITs can become a crucial way that our agency is supported or maintained, as they stand in for typical agential-related constraints and supports when such constraints and supports are absent.

3.3. Social Isolation, Uncertainty, and UITs

If others around us play an important role in helping us express our values and so agency, then when sociality is reduced, we might expect one’s agency will be as well. This is what Doris and Nichols (2012), for instance, have argued. They point to research on solitary confinement where those experiencing such isolation quickly deteriorate, experiencing symptoms of psychopathology that interfere with their agential capacities (434–5; cf. Haney (2003)). Setting aside extreme isolation like that of solitary confinement, more moderate cases of social isolation, like a parent in the postpartum period, are plausibly still instances where one’s agency is negatively impacted. When a new parent—perhaps especially a mother—goes from being surrounded by other moral agents to spending the majority of their day without another agent’s presence and support, social isolation can occur.¹¹ Indeed, reports indicate that social isolation greatly increases during the early postpartum period, when a mother’s time is largely consumed by providing for the basic physical needs of a (non-agential) infant, while also focusing on her own physical recovery and sleep (Barkin et al. 2014; Lee et al. 2019). While not equivalent to solitary confinement experienced by prisoners, if the social account of moral agency is correct, we should expect mothers in the early postpartum period to still be affected.

How might social isolation impact one’s agency? One possible candidate is via increased uncertainty. When surrounded by others, we are given a certain amount of confirmation and confidence—what we are seeing is real, what we believe is true, and what we are doing is right. Roberts and Krueger (2021) explain this connection:

When undertaking decisions and judgments…[there] is psychological labor that is often performed in collaboration with trusted associates whose judgment we trust and with whom we pursue joint projects and collective decision-making...One’s loneliness may thus manifest itself in feelings of…self-doubt (197).

Having others around to collaborate with when we are making decisions, big and small, can help us move around our world with greater certainty and less anxiety. When people are removed, uncertainty rises.

Aside from social isolation, uncertainty can occur for other reasons. In simply moving from a known circumstance to a new and novel one, one will likely experience uncertainty about the ways in which one’s environment can impact them. In explaining his account of agency, Doris (2018) likewise suggests that agency is more likely to be threatened in changing and unstable contexts, “where there are weaker forces for keeping people on the rails” (57). When our circumstances are more rigid, constrained, or predictable, we are better able to act on our values and manifest our agency.

When it comes to unpredictability, becoming a new parent for the first time fits the bill. How one’s new circumstances will impact one’s own thoughts, emotions, and behaviors, is quite unknown. This effect is exacerbated when we are reminded that “any person is capable of forgetting a child in a car” (Rosenblatt 2017). And upon hearing about a mother shaking her baby to death, we find ourselves thinking that “it could have been me or any other of my mum friends…[for she] was just like many of us…What [she] did…is something that I believe many of us are capable of doing” (Anderson 2016). Before being discharged from the hospital, new mothers are often warned that they too may have an urge to shake their own baby. New parents face the possibility that the circumstances they enter into are so wildly different from what they’ve previously known that carrying out such horrible actions now slip into (even the slightest) realm of possibility.

Not only does this all seem intuitively plausible, more empirical approaches which employ a Bayesian inferencing framework also show how entering unpredictable circumstances can increase the occurrence of UITs. Szalisznyó and Silverstein (2021) explain symptoms of OCD, including pathological instances of intrusive thoughts, involve “perceiving the world as more unstable... [and an] overweighting of sensory data...[which] leads to a failure of the use of prior information and less attenuation of sensory noise” (5). They note this “could account for sensory over-responsiveness in OCD, as well as the experience of intrusive thoughts” (ibid). Fradkin et al. (2020) give a similar explanation—uncertainty in one’s current circumstances “leads to greater uncertainty in posterior beliefs… lead[ing] to a failure of the use of prior information…to better attenuate sensory noise.” Information that is usually ignored, skipped over, or kept below the surface of consciousness instead rises to one’s experience (688). And so too with intrusive thoughts: thoughts that would normally not rise to the level of consciousness, or be attenuated, end up manifesting themselves in one’s awareness.¹² Roberts et al. (2021) describe how intrusive experiences—which include UITs but also intrusive images, impulses and bodily sensations—occur when a signal in the brain “gains sufficient biological momentum to breach the threshold of awareness” (250). Such experiences involve “the failure to suspend or attenuate attention to certain kinds of cues” (260). These researchers, too, use a Bayesian hierarchical account to explain when the threshold of awareness is and is not breached, or when intrusive experiences do and do not occur, appealing to a lack of precision of one’s priors and assigning more weight to incoming evidence in an attempt to resolve uncertainty. In cases where one’s priors are uninformative or not applicable, more weight or precision is given to incoming data—and so this information is not attentionally filtered but breaches the threshold of awareness. Likewise, Cochrane and Heaton (2017) give a model of pathological intrusive experiences (as manifesting in OCD) that involves a “hyperactive signal of uncertainty intrud[ing] upon the consciousness of the individual” (185). Possibilities enter one’s mind that cannot be ruled out with certainty:

How do you know you couldn’t harm your child? Do you attempt to do so and then find out you can’t go through with it? Wouldn’t even that experience fail to disconfirm the possibility that you might overcome that constraint if you really tried…? (ibid, 192).

In normal circumstances, these possibilities are too unlikely and are not consciously entertained. What makes these cases of UITs pathological ones is that the person does in fact have the relevant priors or predictions about their situation, such that these possibilities should be filtered out. Yet, plausibly in non-pathological UITs, one’s priors are not suitable to the situation at hand and so are given much less weight. Thus, feelings of uncertainty and thoughts about what is possible (rightfully) cross the threshold of awareness. In other words, the feelings of uncertainty and what might be possible are appropriate, given the uncertainty of one’s environment. Plausibly this is what is happening in the case of UITs arising in new parents: New parents enter new and novel circumstances in which they really do lack sufficient certainty about the harm they are capable of doing.

Additionally, moving from a situation of sociality to one of isolation might not just be a way in which one’s circumstances change and novelty enters in, but isolation itself might also contribute to increased uncertainty and so UITs. As suggested above, social isolation might lead to more self-doubt and a lack of confidence. Furthermore, from an empirical approach, Nenadalová and Řezníček (2025) likewise make use of a Bayesian inferencing framework and find that solitude, or being physically isolated from others, increases internal uncertainty and “otherwise dismissible and implausible [internal and interoceptive] sensations and experiences may arise” (3).

Given these two features of the postpartum period experienced by new parents—namely that it is usually one of (1) increased social isolation and (2) uncertainty—we have a story as to why the frequency of UITs increases for this population. And, these two features likely interact: when one is surrounded by others, these other people often function to hold us to our values and provide a basis of predictable action for ourselves. But when they are absent and we are left to ourselves, uncertainty can arise as to what we might end up doing. This, of course, is exacerbated when we are told that we are capable of doing very bad things that could harm our infant, along with having no prior experience of what these scenarios might be like and how we might respond.