(pictures above are just to be pretty and illuminate what simplified N400 results look like; they are not representative of my experimental work or Perichoresis)

Abstract:  A theory of ‘Perichoresis’ is coined in order to hypothesize the kind of mechanism necessary for a number of our highly productive semantic abilities, such as frame shifting, novel metaphor comprehension, and what is called here ‘deep cognition’.  On a first approximation, Perichoretic mechanisms provide the efficiency necessary for physically realizing the structure and dynamics cued by conventional language as modeled by Fauconnier’s Mental Space theory (1985). On one interpretation of the literature, Perichoretic mechanisms non-redundantly project structure from one Mental Space to another in terms of dynamic mappings captured by traditional principles of ‘presuppositional float.’ Projection is given more specific explanation in this article through analogy to Daniel Dennett’s (1991) discussion on ‘filling in’ and Seana Coulson’s (2006) explicit analogy between Mental Space theory and the various forms of interpolation within the human vision system. The theory of Perichoresis is further expanded through a consideration of ERP N400 component effects (Coulson, 2007) that are produced by nonconventional language, in particular:  novel linguistic metaphor and joke punch lines assumed to cue frame-shifting (Coulson, 2001). Perichoretic mechanisms are tentatively described as the opening of mapping pathways that are ‘deep’, both anatomically and semantically, and they are hypothesized to provide a multi-directional bridge between lexical integration and subsequent elaboration. Perichoresis is also developed to handle conventional language assumed to cue ‘deep cognition’, as in the case of simulation (Bergen, 2007). Perichoretic mechanisms should therefore provide the kind of productive semantic accomplishment common to metaphor, frame-shifting, and conventional deep cognition.  The efficiency of perichoretic mechanisms is explained in terms of their primitive nature, which in turn is explained by the principle of embodied experience grounding semantics. Rather than primarily indexing the ‘difficulty’ of lexical integration, the N400 component is reinterpreted as indexing – in part – primitive processes that yield greater semantic accomplishment per processing cost, in contrast to less productive ‘thin cognition’.

______

Note on the etymology of ‘Perichoresis’:

Perichoresis is a theological term addressing the nature of the Trinity of historic Christian orthodoxy. The term originated in patristic literature and grew to reference the mutual interpenetration or indwelling between each person of the Godhead. In recent years, this doctrine has been reemployed to emphasize the dynamic and relational nature of the orthodox God, counterpoised to the static, impersonal definitions of scholastics, often involving what is analogous ‘feed-forward’ information processes (Coulson, 2006).  The sort of mechanism proposed through use of ‘Perichoresis’ is intended to engage with tradition within philosophy and cognitive science in a similar way.

According to two experiments conducted by Bergen, Lindsay, Matlock, and Naryanan (in Cognitive Science 31, 2007), sentence comprehension yields interference during object categorization when the object is presented in the same location of the subject’s visual field as the event denoted by the sentence (you can locate this article through my link on the right to Bergen’s papers).

In other words, perceiving an object as a particular sort of object is hindered (in terms of processing time) by the pre-triggering of automatic and unconscious mechanisms during normal comprehension of sentences that describe events with the same ‘location’ as the perceived object. This is a reliably produced effect and rigorous controls in the experiments confirm, at the very least, a weak form of simulation theory. So another way of saying this, albeit with a dangerous dose of imprecision and extrapolation, is that the simulated imagery utilized during sentence comprehension is realized by the same mechanisms or neural architecture responsible for visual perception. This can be seen as a linguistic extension of what we have learned about ‘mirror neurons’ found in monkeys, which are structures that activate for both the execution and the perception of a given motor activity or action (Wheeler and Bergen, 2006).

But I introduce these two experiments by way of introduction, since it was the third and the fifth experiments that motivated me to write. (For lack of time and readable space I will not discuss the background experimental details – which is my readers’ loss, since they are brilliant.) These reliably produced interference effects were not found at all during a third experiment, in which concrete actions or events were implied only metaphorically. A fifth experiment was designed to further probe this effect, only to confirm it. Therefore, as a first approximation, linguistic metaphor does not utilize embodied simulation, at least at this fine-grained level of specific location in the visual field.

A fourth experiment had already confirmed that abstract description also failed to produce this interference effect. Sentences were chosen to test specifically for the hypothesized conceptual metaphor MORE IS UP (a metaphor only metaphorically). This suggests that conceptual metaphor likewise does not utilize embodied simulation, at least at such a fine-grained level. However, according to Richardson, et al. (2003) more course-grained, axis specific simulation (e.g. vertical or horizontal planes) does produce the interference found in Bergen et al.’s first two experiments. I am comfortable with Bergen et al.’s suggested explanation that up and down locations are perhaps ‘collapsed’ along the vertical axis during comprehension of more abstract content. (Which calls to mind Fauconnier and Turner’s mechanism of ‘compression’ during conceptual integration).

What I hypothesize from here is that more course-grained simulation is to be found during linguistic metaphor comprehension, whereas the kind of mapping involved in conceptual metaphor does not produce on-line simulation in the same way, if at all. It is tempting – at least for me – to conflate metaphor with conceptual metaphor, but conceptual metaphor is not literally metaphor at all and does not, so I hypothesize, involve mechanisms that can produce the kinds of conscious effects that have motivated simulation theory. I would also think that interpretive cognitive linguistics, in addition to norming studies, might prove illuminating – a thought I had while investigating the sentences used for the fourth experiment, motivated by the conceptual metaphor MORE IS UP. I hypothesize that conceptual metaphors are complex and interpenetrating mappings, such that, ARGUMENT is either WAR, or a GAME OF CHESS (Ritchie), or perhaps both, depending on what precisely it is to be and what it is like to be S at time sequence t.

I would also be interested in integrating these findings with eye tracking experiments during fictive motion (e.g. the chalk board runs along the wall), research suggested by Bergen et al. . I currently see no reason to think that these experiments cannot dovetail with Coulson’s work on the N400 ERP component, which perhaps can be linked to the RTs responsible for the distinction between compatibility (priming) effects and the interference effects discussed in Bergen et al.; relevant RTs appear to be situated between 200 and 1000 ms. Also highly fascinating to me, in consideration of my hypothesis that the boundary between conscious and unconscious simulation and emulation might reveal the underlying structure of consciousness, are the neurological disorders that produce dissociation between axis and location (Bergen et al. 2007, p.755).