The experimental studies on Kaqchikel reviewed so far suggest that the cognitive load during sentence comprehension is primarily determined by grammatical processes operating on linguistic representations, whereas word order selection in sentence production more faithfully reflects conceptual processing at the stage of event apprehension and preverbal message construction. In particular, agent-first orders are likely to be selected over others because of the conceptual saliency of agents. If this conjecture is on the right track, we would expect that the cognitive load during sentence production is higher for SVO sentences than for VOS sentences because the production of a sentence surely includes, as its central part, the construction of linguistic representations, and the grammatical processes involved in this are presumably similar to those involved in the comprehension of a parallel sentence, although there may be some differences. Chapter 10 reports on an experiment to verify this prediction. The results of this experiment support the conclusion that, although Kaqchikel speakers preferentially use the SVO word order because of the saliency of the subject, SVO sentences require more processing resources than VOS sentences both in comprehension and in production.

Chapter 11 discusses some task-dependent effects on word order preference. Different tasks activate different aspects of cognitive processes in the brain. We show that the picture–sentence matching task facilitates the processing of SVO in comparison with the other orders in Kaqchikel, at least partially because of the saliency of agentive concepts accelerating memory retrieval. We also show that Kaqchikel speakers who use Spanish in daily life more tend to process Kaqchikel sentences in the SVO and VSO orders more quickly.

Chapter 14 summarizes and concludes the discussion in this book.

Chapter 6 reports on an ERP experiment with a picture–sentence matching task, in which a picture was presented before the corresponding sentence. The target sentences used in this experiment were the same as those used in the experiment reported in Chapter 5, i.e., transitive sentences with thematically reversible agents and patients, arranged into four word orders: VOS, VSO, SVO, and OVS. The results of this experiment also demonstrated that SVO elicited a P600 compared to VOS, and that VSO elicited a similar posterior positivity, relative to VOS. The results of the two ERP experiments combined clearly indicate that VOS is the syntactically simplest and easiest-to-process word order of the grammatically possible ones in Kaqchikel, which is in line with our previous findings, described in Chapters 3 and 4. In short, Chapters 3 to 6 present data showing that a VOS preference was observed in Kaqchikel sentence comprehension, which provides empirical support for the IGV.

Chapter 5 investigates the time course of the processing of Kaqchikel sentences with alternative word orders. A sentence–picture matching task was employed in an experiment measuring event-related potentials (ERPs). In this experiment, a Kaqchikel sentence was presented aurally through a headset; afterwards, a picture was presented in the center of a screen, either matching the event described by the preceding sentence or not. Upon seeing the picture, the participants were asked to judge whether the picture was congruent with the sentence. The target sentences used in this experiment were all transitive, with thematically reversible agents and patients, arranged into four word orders: VOS, VSO, SVO, and OVS. A late positive ERP component called P600 was used to examine processing loads, as P600 has been found to be elicited by sentences with a filler-gap dependency, reflecting an increased syntactic processing cost. The results of the two experiments demonstrated that SVO elicited a greater positivity (P600) than VOS, and that VSO elicited a similar posterior positivity, relative to VOS. This range of properties follows naturally from the combination of the IGV and the syntactic structures of Kaqchikel transitive sentences given in Chapter 2.

Chapter 7 turns our attention to basic word order in language and natural order of thought. In his seminal work, Greenberg (1963) observed that a vast majority of the world’s languages have one of the SO word orders as their basic word order. It is interesting to note that the distribution is heavily biased even among the three SO orders, with SOV being the most frequent, which indicates that SOV has some special status among the six possible word orders in some sense. Why should this be the case? To address this question, Goldin-Meadow et al. (2008) showed short animations depicting transitive events (e.g., a girl twisting a knob) to speakers of four languages (Chinese, English, Spanish [all SVO], and Turkish [SOV]). The participants were then asked to describe the depicted events by using only their hands, i.e., with gestures. The speakers of all four languages dominantly used the agent–patient–action order in their gestures, regardless of the basic word order of their languages. Goldin-Meadow et al. (2008: 9167) took these results to suggest that the agent–patient–action order reflects the natural sequencing of an event representation and that developing languages use it as the default pattern, thus displaying an SOV word order.

Although previous studies in gesture production have claimed that the agent–patient order is the universal preference when humans think about events and describe them nonverbally, the studies have only assessed languages in which the subject precedes the object in the basic word order (i.e., SO languages). Such limited evidence is not sufficient to conclude that all humans universally perceive the world in the agent–patient order, and it cannot help us disentangle whether the apparent preference for agent–patient sequences is the result of universal cognitive factors or the influence of the word order of SO languages. To disentangle these two possibilities (i.e., the UCV and the IGV), it is crucial to examine a language in which the object precedes the subject in the basic word order. Chapter 8 reports on a gesture production experiment we conducted with Kaqchikel speakers similar to Goldin-Meadow et al. (2008), finding that Kaqchikel speakers dominantly produced agent–patient gestures. Therefore, agent–patient ordering does seem to be a universal preference for event description, which is in line with the UCV as well as the results of previous studies.

Chapter 12 considers the syntactic structure of Kaqchikel sentences in more detail. There are multiple syntactic routes to the VOS order. Different VOS languages may have different syntactic structures. There are two major proposals regarding how Mayan VOS word order is grammatically obtained. We propose in this chapter that Kaqchikel, and possibly Chol as well, derive the VOS order through a right-specifier route, rather than a predicate fronting route.

Chapter 4 compares brain activations in response to Kaqchikel sentences with the VOS and SVO orders, obtained using functional magnetic resonance imaging (fMRI). It is known that the left inferior frontal gyrus (IFG) of the brain exhibits enhanced activation in response to grammatically complex, demanding sentences. The fMRI experiment we conducted with Kaqchikel speakers revealed that cortical activation in the left IFG was significantly higher in response to SVO sentences than VOS sentences, which clearly shows that it is the grammatical features of individual languages, and not universal human cognitive features, that primarily modulate brain activation and determine sentence processing load.

Chapter 13 observes how representative theories of sentence processing costs fare in accounting for the relative difficulties among Kaqchikel sentences with different word orders. The relative processing costs associated with Kaqchikel transitive sentences with the four different word orders are correctly predicted by the Hierarchical Distance Hypothesis, whereas the relative processing costs associated with the corresponding four transitive constructions in Japanese are correctly predicted by the Linear Distance Hypothesis. However, the relative processing costs in these languages are also consistent with the assumption that they are shaped by the combined effects of the three factors: production frequency, linear distance, and hierarchical distance, suggesting the cognitive uniformity of the human parser.