Wednesday, February 15, 2012

Is Priming Necessarily a Threat to Autonomy?

As reviewed by David Nicholson in a previous post, I recently had the privilege of discussing Felsen and Reiner’s “How the neuroscience of decision making informs our conception of autonomy” at a recent Neuroethics Program journal club meeting. The discussion was fruitful and insightful, but as mentioned by David in the aforementioned review post, I think there is a lot more to be said. So, here I am hoping to spark another conversation about Felsen and Reiner’s take on autonomy and neuroscience.

First let me begin be commending Felsen and Reiner for taking on such an ambitious project. They are obviously not the first (or even among the first few) to attempt to outline the relation between the evidence from neuroscience and our capacities for autonomous action (and related capacities such as the capacity to act freely), but they have, without a doubt, taken a much more rigorous approach to the subject matter than most. While many scientists who enter this debate take all measures to stay clear of trying to account for the philosophical conception(s) of autonomy (and related concepts like free will), Felsen and Reiner embrace the challenge of trying to give proper treatment to the relation between a philosophical rigorous account of autonomy and the evidence from neuroscience.

While admitting that the philosophical debates on autonomy are nuanced and ongoing, Felsen and Reiner identify three broad principles that, in their view, are relatively standard to most philosophical accounts of autonomy. According to Felsen and Reiner, in order for a decision to be autonomous it must be:
  1. Consistent with the individual's higher-order desires.
  2. A product of rational processes.
  3. Not unduly influenced by external factors beyond the individual's control.
Felsen and Reiner go on to offer evidence for or against each of these requirements of autonomy, finally concluding that the “neuroscience of decision making is consistent with the standard model of autonomy” but only after some qualifications to the standard model. The requirement that they see most difficult to meet – and is the requirement that they see as most likely needing revision – is the requirement of not being unduly influenced by external factors.

Felson and Reiner correctly note that, in order to act autonomously, we do not need to be completely free from external influences. We just need to be free from undue influences beyond the control of the individual. Felson and Reiner offer priming (and priming-like effects) as what they take to be an uncontroversial example of the sorts of external influences that should be seen as “undue” and “beyond the control of the individual.” At first glance, priming seems to be the perfect example of an autonomy-compromising phenomenon. After all, priming occurs automatically, priming occurs unconsciously, and priming is generally thought to be beyond an individual’s control. However, I want to suggest that priming is not necessarily a threat to autonomy. In fact, I want to suggest something (seemingly) even more radical: Priming can often help facilitate our ability to act autonomously.

To appreciate this claim, let us first consider the fact that we are constantly bombarded with a large array of information from the environment. At any given moment, we are only consciously aware of a minutia of the information that our systems are processing. This means that much of the information that our systems are processing is being processed unconsciously and automatically. All this information – including, if not especially, the unconscious and automatically processed information – provides cues regarding how to proceed at any given moment. These cues effectively “prime” us toward the selection of certain behaviors over other behaviors. The priming provided by the automatically and unconsciously processed cues contribute to making us much more effective and efficient at navigating our world (including making us much more effective and efficient at making goal-directed decisions). Priming provides efficiency (and often effectiveness) in our decision-making that may not otherwise be obtainable. And I think some degree of efficiency and effectiveness is required to be an autonomous agent. (Imagine being burdened by the constant requirement of being able to rely only on consciously processed information in order to make decisions. It would be paralyzing, not liberating.) So, in this sense, priming (and priming-like effects) may actually be able to help facilitate autonomous actions.

Of course, not all priming is equal. Priming that pushes us toward the selection of behaviors that are not consistent with our higher-order desires would be an example of priming undermining autonomy. But even in this case, it is not the priming per se that poses a problem for autonomy; rather it is the fact that there is inconsistency between the effects of the priming and our higher-order desires. These sorts of autonomy-compromising conditions are already noted as autonomy-compromising via Felsen and Reiner’s first requirement of autonomy. So, priming only really becomes an “undue” external influence when it violates one of the other requirements of autonomy, not because priming itself is necessarily an “undue” external influence.



Want to cite this post?
Shepard, J. (2012). Is Priming Necessarily a Threat to Autonomy? The Neuroethics Blog. Retrieved on , from http://www.theneuroethicsblog.com/2012/02/is-priming-necessarily-threat-to.html


--Jason Shepard
Emory Neuroethics Scholars Program Fellow

4 comments:

Kristina Gupta said...

Hi Jason,

I suppose I would be interested in exploring further Felsen and Reiner's rather short discussion of social and cultural influences. They write: "few (if any) decisions are entirely independent of the individual's family, culture, and other social groups." According to their model, external influences which are consciously incorporated aren't a threat to autonomy, but unconscious influences are. But their discussion of unconscious family, cultural, and social influences is rather anemic. For example, they write: "childhood education, generally considered to be a positive feature of healthy societies, entails the influencing of beliefs and actions." This seems to be a rather rosy view of public education, which can also be considered a vehicle for instilling jingoism and conformism. I like the discussion of authenticity in Bolt and Schermer's article "Psychopharmaceutical Enhancers: Enhancing Identity" (2009), particularly their discussion of the 1950s middle-class housewife. A 1950s housewife, in deciding to conform to traditional gender norms, may be following her "first-order" desires (which are themselves influenced by social factors), but there may still be a sense in which she is not living an authentic life. How does authenticity relate to autonomy?

Kristina

Jason Shepard said...

Hi Kristina,

Schectman's example of the 1950's housewife not living an authentic life has little to do with conforming to the expectations of the cultural milieu in which you belong, and a lot more to do with what the identity conditions of one's "true self" (or "authentic self" or in Schectman's preferred vocabulary one's "true nature").

For Schectman (at least Schectman as presented in Bolt and Schermer), for one to be acting freely, one must act in a manner consistent with one's true self (with one's true nature). This claim is consistent with a Frankfurtian account of free action. However, where Schectman and Frankfurt differ is that Frankfurt claims that acting in accordance with wholeheartedly endorsed second-order desires is sufficient for saying that one is acting in accordance with one's true self. Schectman says, "Not so fast!" and gives the example of a particular 1950's housewife who has a very particular "nature." (Again, this example is not meant to be an example about whether any 1950's housewife can act freely.)

In the particular example, the 1950's housewife has conflicting first-order desires. She desires to be a good, feminine housewife, but she also desires a job, to be politically active, and so forth. She really wants her desires related to being a good, feminine housewife to be the desires that move her action. Supposedly, according to the example, she wholeheartedly and without reservation wants the desires related to being a good, feminine housewife to be the desires that move her action.

The intuition here is that these wholehearted second-order desires do not really reflect this particular 1950's housewife's "true self." Rather, this particular 1950's housewife is using her will in order to go against her true self. Thus, she is not really "acting freely" since the first-order desires that move her actions are not the first-order desires that best reflect her "true self" (they do not reflect her "nature", her "authentic self").

But the intuition only works here, if we think at some level the housewife really wants the desires contrary to being a good, stereotypical housewife to be the desires to move her. But this seems to betray the fact that the housewife does not truly wholeheartedly endorse the values of the good housewife (because deep down she really endorses values that are inconsistent with the values of the good housewife; thus, she is actually very divided).

______________________

However, I think your question was really about whether 1950's housewives, in general, can act authentically when they are acting in accordance with gender stereotypes. In short, I think the answer has to be "yes" (or else is it doubtful that anyone can be said with confidence to be acting authentically).

Peter B. Reiner said...

Jason –
Gidon and I are flattered – and delighted! - that our paper has evinced so much discussion and thought. Nothing could please us more. Moreover, we are even more enchanted with the fact that you take up the intellectual cudgel and challenge our ideas – in the final analysis, I think we are all interested in exploring these issues with an eye towards clarity, and certainly the discussion on this blog helps.
That being said, I do wish to take issue with the main point of your post. You say,
“The priming provided by the automatically and unconsciously processed cues contribute to making us much more effective and efficient at navigating our world (including making us much more effective and efficient at making goal-directed decisions). Priming provides efficiency (and often effectiveness) in our decision-making that may not otherwise be obtainable. And I think some degree of efficiency and effectiveness is required to be an autonomous agent. (Imagine being burdened by the constant requirement of being able to rely only on consciously processed information in order to make decisions. It would be paralyzing, not liberating.) So, in this sense, priming (and priming-like effects) may actually be able to help facilitate autonomous actions.

I am reasonably confident that the first two sentences of your comment are quite correct. But I think you rest on much shakier ground when you conflate efficiency and autonomy. In the strictest view, an autonomous decision is one that I have made based upon information that I have and, importantly, is not intruded upon by outside forces. Whether that process is efficient or not is not, in my view, part of the equation. It may be desirable to be efficient, but that efficiency does not contribute to autonomous decision making per se.

Jason Shepard said...

Hi Peter,

Thank you so much for your thoughtful comments. I was really hoping someone would call me out on the very particular step that you are challenging. (The step is both central to my point and easily the most controversial step that I make.)

This notion of relating efficiency with autonomy on its face seems counter-intuitive (perhaps just wrong!). But what I was hoping to latch onto here is that some degree of efficiency is required for autonomy. If one's definition of autonomous decision making requires that all the inputs that go into conscious decision making also need to be conscious in order for any given decision to be considered fully and truly autonomous, this seems a bit problematic for two reasons.

The first problem is that, if this definition is correct, then it seems likely that none of our decisions are (fully) autonomous. Now, with this problem in mind, you may come to one of two conclusions: The definition is right and thus there are no (fully) autonomous actions. Or, there are fully autonomous actions, and thus this definition may need to be revised. (If I am not mistaken, this latter conclusion is the one that you favor.)

The second problem is that, given the way our conscious system(s) operate, if all inputs into a conscious decision were also required to be conscious, then this requirement would obliterate efficiency to the point that any "autonomous" decision making process would be paralyzing. (Would we even be able to make relatively simple decisions under this model?) So, this leads me to think that some degree of efficiency can actually promote autonomy (and may actually be required for autonomy). And it seems like unconsciously processed external stimuli may actually be the sorts of things that can grant us the bit of efficiency that may promote (would be required of?) autonomous decision making.

Now, of course, sometimes efficiency (or efficiency-promoting processes) obtrudes upon our ability to act autonomously.

With this view in mind, we should be careful of saying that some type of efficiency-promoting mechanism automatically undermines autonomy. Rather, I think we should be asking: Under what sorts of conditions would efficiency-promoting processes undermine our ability to act autonomously?

And whether you buy into my view or not, I still think the "conditions" question is more informative and better posed than the "types" question. That is, asking and trying to answer the question, "Under what conditions do certain processes undermine autonomy?" is more informative and better posed than asking and answering "What types of processes undermine autonomy?"