Read Aloud the Text Content

This audio was created by Woord's Text to Speech service by content creators from all around the world.

Text Content or SSML code:

A more transparent relationship between sentence elements would be present if the above sentence were “The girl who chased the boy tripped over a cat,” since the Wh- movement is vacuous, and the assignment of a thematic role to “The girl” can be achieved based simply on canonical word order—NP1 = Agent, Verb, NP2 = Theme (i.e., Subject NP is assigned thematic role of “Agent” by next Verb). On the other hand, the relationship between the Subject NP and the matrix verb “tripped” in both of the above sentences is not transparent due to the intervening embedded clause. Namely, in order to assign Agent to “The girl”, it is necessary to link this NP via the syntax to the matrix verb “tripped.” If syntactic relations were not computed, then the canonical interpretation of the matrix clause would posit “the boy” as the Agent of “tripped” (i.e., because the closest NP preceding the verb is “the boy”). However, as was the case for the preceding sentence, computing such syntactic relations makes processing the matrix clause in this sentence more difficult. In particular, in order to assign a thematic role to “The girl,” it is necessary to hold this Subject NP in memory while the embedded clause is processed and until the matrix verb “tripped” is encountered. This holding and concurrent processing places demands on working memory. An alternative or complementary explanation of agrammatic symptoms makes reference to the timing of operations when computing information in a sentence (Lenneberg, 1967; Kolk, 1995--see timing figure below). Some research shows that lesions to Broca's area result in slower activation of lexical information, which leads to slower processing overall (Blumstein & Milberg, 2000; Swinney et al., 1999). The end result is that information is not available when needed, and therefore, is left out or not integrated with other sentence information. For example, in the sentence “The girl who the boy chased tripped over a cat,” slower activation of the NP “the girl” could result in it not being sufficiently activated (or re-activated at the trace) to receive a thematic role when the verb “chased” is encountered. Or, the combined effect of slower activation of “the girl” and “the boy” may mean that processing resources (energy) that could have been used to process “chased” would not be available due to the continued activation of the two NPs. In either of these scenarios, the reduced energy available to activate lexical information leads to insufficient resources for completing both lexical and syntactic operations and establishing thematic relations among different elements in the sentence. The theoretical accounts of processing limitations and timing disturbances can be unified in Caplan's (2006) or Carpenter and Just’s “Resource Reduction” explanation of aphasia (e.g., Haarmann, et al., 1997). Just and Carpenter propose a trade off in storage and processing such that when a person’s resources are limited they may be preferentially allocated to lexical processing which may cause syntactic processing to suffer; or if directed to syntactic processing, then lexical processing suffers. The latter appears to have been the case in the Burkhardt et al. (2003) study where complex sentences (with Wh- movement) resulted in protracted reactivation of the lexical filler (see below). A reduction in processing power due to a brain insult (or due to external constraints in the case of individuals with intact brains) forces the language processing system to make choices about where it will direct available cognitive processing power. If more of this energy is allocated for activating lexical items, then these items may be produced or comprehended, but not in relation to other elements in the sentence. On the other hand, if energy is directed to establishing thematic roles, then all of the available processing power may be used up in getting just one thematic role computed, leaving other roles and relationships in the sentence unassigned. In other words, in sentence processing, focusing on storage (remembering each argument and verb) can lead to slower processing and inability to assign thematic roles; alternatively, focusing on establishing thematic roles (processing) may lead to one or more arguments being lost (from storage). On-line Experimental Evidence Slow lexical activation in Broca’s aphasia (Aydelott Utman et al., 2001; Prather et al., 1997) Delayed gap-filler reactivation in sentences with Wh-movement (Burkhardt et al., 2003; Swinney et al., 1999) Burkhardt et al. (2003): tested the prediction that syntax-dependent effects such as ‘‘gap-filling’’ are observable but in a delayed fashion in Broca's aphasia. Method: Cross-modal priming experiment Auditory sentence presentation with visually presented word/nonword probe before, at, or after a trace/gap Subjects responded with a lexical decision to probe The probes were either related or unrelated to the antecedent ("cheese"), or a nonword “The kid loved the cheesej whichj/i the brand new microwave (P1) melted ti (P2) yesterday (P3) afternoon while the entire family was watching TV.” Probes: cheddar (related “prime”), album (unrelated), gorpul (nonword) Results: Broca’s participants only showed related-probe priming of antecedent at P3 (not P2, where normal participants do), suggesting they experienced slower reactivation of “cheese” at trace/gap. Further evidence to support this interpretation comes from Swinney et al. (1999), who reported that Broca’s aphasic participants re-activated the Wh-filler at the trace site when speech rate was slowed. That is, when additional time was provided for processing a sentence, lexical (re)activation proceeded more normally (see also Love et al. 2008). Paraphasia & Paragrammatism Both linguistic and psycholinguistic explanations have been offered for fluent aphasic individuals’ paraphasic and paragrammatic production and difficulties with language comprehension. Linguistic accounts have attributed errors in phoneme, morpheme, or word selection, and in word comprehension, to impairments to input and output morphological, lexical-phonological and/or lexical-semantic representations. That is, substitutions of phonemes (e.g., telephone-->lelephone), morphemes (not so much the case in English as in more morphologically rich languages such as German or Italian), and words (e.g., harp-->harmonica) in production arise because of faulty phonological, morphological, and lexical representations and/or selection. In comprehension, the same domains of representation are considered to be impoverished due to the brain lesion, leading to unreliable phoneme, morpheme, and word mapping, and ultimately, disrupted comprehension. Psycholinguistic Account of Paraphasia and Paragrammatism A psycholinguistic account does not dispute the linguistic elements implicated in paraphasic/grammatic behaviour. However, instead of attributing these symptoms to loss of linguistic competence, it takes into account cognitive processing mechanisms that underlie the language problems. For example, in a connectionist network, linguistic representations are represented as nodes in the network, and these nodes are selected based on the amount of activation they receive in conjunction with their inhibition of competing nodes at a particular level of processing (e.g., phonemes, morphemes). The latter process is what appears to be disrupted in fluent aphasia. Namely, whereas activation of linguistic elements is not problematic, keeping this activation “in check” is, in the sense that irrelevant elements should not be activated to the same degree as relevant (target) elements. Hence, failure to inhibit competing phonemes, morphemes, and words leads to paraphasic/grammatic behaviours that resemble the target in form and/or meaning. A complementary explanation of paraphasic/grammatic symptoms has been construed in terms of the declarative-procedural distinction (from research on memory). According to Ullman (2004), the frontal (Broca’s) and posterior (Wernicke’s) aphasic syndromes correspond to cortical support for procedural and declarative information processing systems. The procedural system handles highly automatized processing, whereas the declarative system engages in more controlled conscious information processing. Thus, in Broca’s aphasia automatic processing is affected leading to slower activation and timing disturbances (more effortful and controlled processing). In Wernicke’s aphasia, automatic processing is preserved, whereas posterior-supported inhibition (controlled) processing takes a hit. Conclusion: Cognitive processing resource accounts of aphasic disturbances provide an elegant explanation of a number of different agrammatic and paraphasic/paragrammatic symptoms. Furthermore, they can explain the correspondences in performance between children, aphasic individuals, and normal adults. Processing limitations also can account for the fact that aphasic individuals’ performance is variable—sometimes correct forms are used or understood, other times not. Such variability can be attributed to fluctuations in overall energy available at any point in time, and to the different ways in which limited resources can be allocated in performing a task. Altogether, the evidence points not to a primarily linguistic representational loss but rather to impaired operations on those representations leading to lack of activation or inhibition of certain elements and/or inadequate integration of elements over time. Additionally, one must question whether previously identified cortical regions as “language” centres are not in fact more adequately characterized as “processing” centres that vary along dimensions of automaticity and inhibition.