Despite a rich, comprehensive literature on aspect in linguistics, its role in reporting past events is still poorly understood. In particular, little is known about how aspect biases the way people formulate thoughts and generate utterances about dangerous or emotionally charged events they have witnessed firsthand. Of special interest here was aspectual framing. We used a naturalistic task to explore how people would spontaneously talk about car accidents in response to an open-ended question that included imperfective or perfective aspect. Participants first viewed videotaped car accidents, and then were asked to explain what happened or what was happening. As predicted, this instructional manipulation resulted in consistent differences as to how actions were reported. Mainly, imperfective framing yielded more action details, evidenced by more motion verbs, more reckless driving language, and more iconic gestures. We also found that imperfective framing resulted in fewer non-motion verbs and fewer beat gestures. No differences were observed for number of words, number of gestures in the two conditions. Neither was there a difference in imperfective and perfective forms generated.
The results of our experiment suggest that imperfective framing led people to pay more attention to action details in formulating their descriptions. One feasible explanation for this resides in perceptual simulation, an embodied, perceptually-grounded mechanism that drives much of our everyday reasoning. Simply stated, simulations are re-activations of patterns that are anchored in past perceptual and motor experiences (see Barsalou 1999, Glenberg 1997). A rapidly growing body of behavioral studies has shown that simulations are involved in many facets of everyday thought, including concept formation (Barsalou 1999), reasoning about physics (Schwartz and Black 1999), reasoning about spatial relations (Richardson, Spivey, Barsalou and McRae 2003, Spivey and Geng 2001), and conceptualizing abstract domains, such as time (Boroditsky and Ramscar 2002, Matlock, Ramscar and Boroditsky 2005). In addition, neuroscientific research provides substantial evidence to suggest that people readily simulate action (Gallese 2005, Jeanerrod 1996). For instance, when people observe others performing an action (e.g. watch a person grasp an object), activation in their motor areas unfolds in a manner that is consistent with how it would occur if they were performing the very action themselves (Rizzolatti and Sinigaglia 2008). Similarly, when people view static images of humans in motion (e.g. look at a photograph of a person who appears to be throwing a ball), motion perception areas are activated, and they simulate the experience of seeing the action (Kourtzi and Kanwisher 2000). And psycholinguistics research supports the idea that simulation figures into linguistic processing. It is known, for example, that simulation is involved in understanding literal language (Glenberg 1997, Pecher and Zwaan 2005) as well as non-literal language, including conceptual metaphor (Gibbs 2006b, Gibbs and Matlock 2008) and fictive motion (Matlock 2004, Richardson and Matlock 2007). Despite mounting evidence for this, many language theorists continue to maintain the position that linguistic processes do not include simulation. Rather, linguistic processes are typically characterized in terms of specialized modules, and thus, as largely blind to perceptual and motoric information, including simulated versions of perception and action. Meaning is viewed as a byproduct of syntactic form, and in some cases, is achieved through executive control (see Fodor 1975, Jackendoff 2002). (For comprehensive discussion of anti-embodied approaches to language, see Barsalou 2008, Gallese and Lakoff 2005, Gibbs 2006a, and Pecher and Zwaan 2005).
Recent work has begun to explore aspect and simulation. Some studies compare the way actions are conceptualized with imperfective aspect versus perfective aspect (see Matlock 2010, 2011). Thus far the results indicate that imperfective aspect affords rich simulations of events by drawing attention to details of events as they unfold in time, and that perfective aspect has less potential for rich simulation of action details because it emphasizes the completion of an entire event. (See Bergen and Wheeler 2010, Huette, Winter, Matlock and Spivey 2012, Madden and Therriault 2009, Narayanan 1995, for compatible research.) Critically, in the results reported here, imperfective framing led to the encoding of more action per situation than did perfective framing. In thinking about what was happening just moments earlier in a video they viewed, people "played back" many rich action details. In simulating these details and formulating their descriptions, they provided reliably more motion verbs, reckless language, and iconic gestures. In contrast, perfective framing resulted in weaker, less vivid simulations, which gave rise to fewer action details in speech and gesture, but more time for non-motion verbs and beat gestures. (For supporting work on simulation and gesture, see Hostetter and Alibali 2008.) Another, not incompatible explanation for our results is that people took an internal perspective with imperfective framing and an external perspective with perfective framing (see Madden and Zwaan 2003, McNeill, 2003). An internal perspective would mean greater access to action details than an external perspective would, and this could result in more motion verbs, more reckless driving phrases, and more iconic gestures.
This work on aspectual framing contributes new insights to research on language and eyewitness testimony. To date, much of the work on language in the courtroom has focused on lexical content, and ignored grammatical content (see Loftus and Palmer 1974). Based on the findings reported here, it is reasonable to assume that aspectual framing may be useful in the courtroom. Attorneys, for example, could ask questions with imperfective aspect to implicate criminal intent or emphasize the magnitude of immoral acts. This could potentially help sway jurors or judges, and result in considerably longer jail sentences and larger fines. Support for this line of reasoning comes from related work on the influence of aspectual framing in political messages. In Fausey and Matlock (2011), participants read a brief passage about a senator who exhibited undesirable behavior in the past, and then answered questions, including whether they thought the senator would be re-elected and about their degree of confidence about their decision. When the senator's actions were described using imperfective aspect, such as was asking hush money from a prominen£ cons£i£uen£, participants were more confident that he would not be re-elected than when his actions were described with perfective aspect, such as £ook hush money from a prominen£ cons£i£uen£. Imperfective aspect also resulted in higher dollar estimates in responses about the amount of hush money taken.
Our findings are consonant with cognitive linguists' claims about the semantics of aspect and event construal. Lakoff (1987) argues that utterances are not the concatenations of fixed or autonomous words that "live" in an idealized monolithic lexicon (see also Clark 1997). Rather, they are grounded in human perceptual and motoric patterns of experience. On this view, aspect, and other linguistic systems associated with events naturally emerge from these embodied interactions. From this, it follows that imperfective aspect reflects the way humans view and enact actions that are ongoing, repeated, or habitual. Similarly, Langacker (1987, 1990) argues that the semantic import of grammatical systems, including grammatical aspect, is anchored in perceptual and cognitive experience. In this case, the distinction between imperfective and perfective aspect is motivated by differences in dynamic conceptualization, the way conceptual structure develops during linguistic processing. Imperfective framing resonates with Langacker's (1987) sequential scanning, in which component states of a situation are scanned serially, and perfective framing with summary scanning, in which component states are scanned in a single gestalt (see Broccias & Hollmann 2007 for insightful discussion). In a similar way, Talmy (1985, 2000) views language as a cognitive system that draws on other cognitive systems, especially visual perception. He argues for a common conceptual capacity that unifies seemingly disparate realizations of linguistic form (e.g. lexical versus grammatical) within a language and across multiple languages. On Talmy's view, the basic distinction between imperfective versus perfective is in keeping with that of mass versus count nouns (see also Langacker 2000). From this, it makes sense that our participants conceptualized and articulated more action with imperfective framing than with perfective framing.
Our results are also in line with some non-cognitive linguistic work on aspect, including the more foundational semantic characterizations of imperfective and perfective aspect, useful in typological or comparative analyses. For example, Comrie (1976) argues that imperfective aspect emphasizes an internal perspective of a situation, and that perfective aspect emphasizes a global perspective. Given this, it follows that when people take an internal perspective, they will devote much attention to action details, but if they take a more global perspective, they will not. It is unclear, however, how our results can inform or resonate with generative linguistic approaches to aspect and event descriptions. Work on the semantics of aspect in natural discourse is limited even though there is some research on aspectual shifts (e.g. Smith 1991).
The current study sheds new light on the role of aspect in natural discourse, specifically, its power to influence the way past events are reported. Many issues remain. It would useful to run a similar experiment on aspectual framing with speakers of languages that have notably different aspectual systems, such as Finnish, Spanish, Indonesian, and Russian. This would lead to an even better understanding of how aspectual framing influences retelling, and how the effect may be across languages. It would be informative to investigate aspectual framing in even more natural situations, for example, interactions with two or more people engaged in a problem-solving task (e.g. Clark and Krych 2004). It may also be informative to conduct studies that pinpoint when and how gestures occur relative to speech in the context of aspectual framing (e.g. path gesture with motion verb). The duration of gesture strokes is known to be longer in imperfective descriptions than in perfective descriptions (Duncan 2002), but more work could lead to an even better understanding of the temporal dynamics of aspect in gesture. It could also be informative to run studies on aspectual framing and gesture with humans and avatars in interactive virtual learning environments (see Huang, Matthews, Matlock and Kallmann 2011, Huette, Huang, Kallmann, Matlock and Matthews 2011, for research on motion capture and gesture). It would also be worthwhile to assess the utility of aspectual framing in a variety of social domains, including doctor-patient interactions and teacher-student interactions.
Far more work can be done on aspectual framing. For now, however, we can say that aspect was playing, is playing, and will continue to be playing a vital role in shaping how we think about and talk about everyday events.