Neuroscience, Self-Understanding, and Narrative Truth

Abstract

Recent evidence from the neurosciences and cognitive sciences provides some support for a narrative theory of self-understanding. However, it also suggests that narrative self-understanding is unlikely to be accurate, and challenges its claims to truth. This article examines a range of this empirical evidence, explaining how it supports a narrative theory of self-understanding while raising questions of these narrative's accuracy and veridicality. I argue that this evidence does not provide sufficient reason to dismiss the possibility of truth in narrative self-understanding. Challenges to the possibility of attaining true, accurate self-knowledge through a self-narrative have previously been made on the basis of the epistemological features of narrative. I show how the empirical evidence is consistent with the epistemological concerns, and provide three ways to defend the notion of narrative truth. I also aim to show that neuroethical discussions of self-understanding would benefit from further engagement with the philosophical literature on narrative truth.

Keywords:

• memory
• neuroethics
• persons
• self
• truth

Acknowledgments

Work on this article was supported by ARC Discovery Grant Neuroethics: The practical and the philosophical. I thank Jeanette Kennett, Jacqui Poltera, James Franklin, and Andrew Walker for reading and commenting on drafts.

Notes

Some of these authors speak more directly, connecting neuroscientific evidence to narrative theories of the self. I focus on self-understanding in this article, as I will not be addressing the metaphysical issues raised by claims about narrative selfhood.

The exact nature of these connections is the subject of some discussion (see Carroll 2001; Christman 2004; Currie 2007; Schechtman 1996; Velleman 2003). I assume here they would include causal, teleological, and other explanatory connections.

A similar role is played by Korsgaard's (1996) notion of “practical identity,” a guiding self-image that includes one's reasons for action.

A detailed discussion of this feature of autobiographical memory is provided by Schechtman (1994), discussed further, later in this article. The basic idea here is likely familiar to most of us, however.

It is important to note that the processes of self-interpretation briefly described here are intersubjective: Our self-narratives—and the norms we draw on in creating them—are constructed in constant interaction with others, and so are dependent on social and cultural context. Although I do not have space to do proper justice to this aspect of narrative theories here, it does have important implications for scientists and psychologists working in these areas, and some of the connections are noted below. For detailed work on narrative and intersubjectivity see, for instance, Ricoeur (1970; 1992), Nelson (2002), Atkins (2005, especially 45–56), and Gallagher (2009).

Again, as well as the interpretations themselves being dependent on social and cultural context, precisely how a person enacts these interpretations will differ according to context—so that a trait like “being organized” may be emphasized in situations where social norms encourage it (e.g., at work) but not emphasized in other domains.

For versions of this claim (of varying strengths) see Currie (2006, 314); Vice (2003, 95–96), and Strawson (2004, 440).

This is consistent with a number of social psychology studies that show that memories tend to be skewed in ways related to our desires and our self-conceptions, discussed later.

For a thorough discussion of different confabulation syndromes see Hirstein (2005). Another way in which confabulation syndromes might be relevant to narrative theories of self-understanding is in the relentlessness with which some patients confabulate (e.g., Sacks 1985, 103–110). Some have interpreted Korsakoff's syndrome as the patient's attempt to fill gaps in memory (Hirstein 2005, 17–18), which is suggestive of a certain need to have some story in play throughout our lives—false or otherwise.

This theory is evidenced by a double dissociation. Damage only to areas associated with memory systems may cause memory loss without confabulation; damage to some orbitofrontal regions or connections may cause inability to distinguish fantasy from reality (Hirstein 2005, 59) or lack of epistemic inhibition (2005, 81), without confabulation. For instance, Hirstein's review of studies of Korsakoff's syndrome reveals that patients often have damage within a “circuit” composed of the thalamic mediodorsal nucleus, amygdala, and orbitofrontal cortex, coupled with damage to the mamillary bodies, medial temporal lobe or hippocampal area (Hirstein 2005, 48–49). The thalamic mediodorsal nucleus is thought to play a role in blocking or suppressing inappropriate thoughts and emotions (2005, 52), and damage to the orbitofrontal cortex has also been associated with other forms of “epistemic disinhibition” such as compulsive lying (Hirstein 2005, 81). The mamillary bodies and hippocampal systems are associated with accessing memories. Hirstein also finds a double dissociation for a range of other confabulation syndromes.

See Gazzaniga (1995) for an explanation of methods used to convey information to only one hemisphere.

This is loosely in line with Hirstein's hypothesis, although confabulation syndromes were not always obviously lateralized. With regard to unawareness of illness, some have posited that lesions in the right hemisphere are more likely to cause confabulation than those in the left (Hirstein 2005, 143; but see also 176).

For more on the sometimes dramatic prevalence and extent of everyday cognitive errors, see Simons and Chabris (1999; 2009). This work suggests that the sorts of cognitive errors we tend to make with regard to our self-understanding are instances of broader tendencies toward certain kinds of cognitive errors.

As these interpretations are evaluative, these studies do not necessarily mean we are all therefore mistaken. Rather, this evidence indicates a tendency toward certain biases in our self-interpretations.

These ideas also have implications for attempts to effect behavioral change among some populations. For an example, see Hochstetler et al. (2010) on violence and self-conceptions in methods for rehabilitating violent offenders.

For instance, see Benabou and Tirole (2003) and Akin-Little et al. (2004).

Arguably, a confabulated reason of this sort could count as a real reason for action, if we think that some first-person statements are immune to error, or indeed that these subjects may have also felt like a soda or found the experimenters “too much.” If this is right, the error here would be that subjects misidentified these reasons as the direct causes of their action, rather than that the reasons themselves were false.

Bickle targets the fictionality of what he calls the “self-in-control” in particular in his discussion of the fictionality of our internal narratives. He does also acknowledge that the evidence currently available leaves discussion of some of these issues necessarily speculative (Bickle 2003, 204–205). The discussion in the following section provides support for moderating such speculation.

The constraint is also meant to deflect the implication of Schechtman's version of the narrative view, on which we constitute ourselves as persons through enacting our narrative self-conception, that any narrative no matter how inaccurate would be made true by being enacted. I discuss this later.

Although the reality constraint is somewhat vague, it is intentionally so, because Schechtman wishes to recognize that the boundaries between acceptable and unacceptable inaccuracies in self-narratives are blurry, and that narrative personhood is a matter of degree. Extended examples to clarify these statements defining the reality constraint are given in Schechtman (1996, 121–130). Note that I am not endorsing the reality constraint here, rather demonstrating that there are ways of differentiating problematic from acceptable self-narratives that could be further refined. The reasons that any such distinction will however need to allow for some blurriness about which inaccuracies are acceptable are discussed further, later in this section.

An issue I shall not attempt to deal with here is that this may lead to considering “reality” a matter of consent. For instance, in cases of political oppression, the oppressed may come to believe they are by nature inferior, but this does not make it true.

Baron (1989) makes a similar point in her work on self-deception as arising from the same structures of self-understanding that enable good self-knowledge.

These ideas are developed in Ricoeur's work on psychoanalysis and are intricately linked to his understanding of the psychoanalytic relationship as the source of psychoanalytic “facts.” This approach relates to the notion of self-narratives as intersubjectively created. See Ricoeur (1970) and McCarthy (2007, 201–214).

The way that the truth about some event can depend on what happens later is demonstrated in much more detail in Lloyd's discussion of Sartre's autobiographical Words (Lloyd 2008). For another way to describe how truth can depend on what happens later, see also Jones's development of the notion of “trajectory-dependent properties” (Jones 2008).

For a more developed example in which we can question whether there is some final fact of the matter corresponding to self-interpretations, see van Fraassen's (1989) discussion of self-deception.