Agrammatic comprehension represents one of the most intellectually compelling puzzles in cognitive neuroscience and clinical aphasiology, challenging classical dichotomies of expressive versus receptive language impairment. Far from experiencing a generalized breakdown in linguistic understanding, individuals with this condition exhibit selective deficits in deciphering non-canonical syntactic structures while retaining intact lexical and semantic processing. Investigating this phenomenon not only illuminates the neural architecture of syntactic parsing in the human brain but also directly transforms neurorehabilitation practices for individuals recovering from focal neurological trauma.
Agrammatic Comprehension
1. Concise Definition
Agrammatic comprehension is a specialized neurogenic language impairment characterized by the selective inability to understand sentences whose thematic roles depend strictly on syntactic structure, syntactic movement, or grammatical functional morphology, while lexical-semantic comprehension remains largely preserved. Individuals exhibiting this profile typically understand canonical, active-voice constructions (e.g., "The dog chased the cat") via real-world semantic plausibility or standard agent-first heuristics, yet perform at chance levels on structurally complex or non-canonical sentences (e.g., object-relative clauses and passive constructions) where thematic role assignment necessitates algorithmic syntactic parsing.
First recognized systematically within the framework of Broca’s aphasia, agrammatic comprehension demonstrates that the traditional dichotomy dividing aphasic syndromes strictly into expressive motor deficits and receptive sensory deficits is clinically and anatomically invalid. Instead, agrammatic comprehension reflects a central disruption in the computation, retrieval, or maintenance of hierarchical syntactic dependencies, positioning it as a foundational construct in modern cognitive neuropsychology and neurolinguistics.
2. Etymology & Linguistic Origin
The term derives etymologically from the Classical Greek root gramma (γράμμα), meaning "letter," "writing," or "that which is drawn," which evolved into grammatikē tekhnē, the art of grammar. The privative prefix a- (α-), denoting absence, negation, or deprivation, combines with this root to form agrammatos (αγράμματος), historically signifying "unlettered" or "illiterate." The term agrammatism was first introduced into the neurological literature in the late nineteenth century by Adolf Kussmaul (1877) and later refined by Arnold Pick (1913) to describe telegraphic, simplified speech output devoid of function words and morphological inflections.
The specific compound construct agrammatic comprehension emerged during the mid-1970s through the landmark neuropsychological investigations of Alfonso Caramazza, Edgar Zurif, and their colleagues. Prior to this inflection point, comprehension deficits were classified primarily under receptive aphasic profiles such as Wernicke’s aphasia. Zurif and Caramazza demonstrated that patients diagnosed with classic Broca’s aphasia—long presumed to possess pristine comprehension—displayed severe, systematic comprehension failures when forced to rely entirely upon grammatical morphology and syntactic word order. Consequently, the descriptor shifted from being purely an expressive diagnostic marker to denoting a central, bidirectional syntactic processing deficit.
3. Pronunciation & Grammatical Form
The term is pronounced phonetically as /eɪ.ɡrəˈmæt.ɪk kɒm.prɪˈhɛn.ʃən/ in British English and /eɪ.ɡrəˈmæt.ɪk kɑːm.prɪˈhɛn.ʃən/ in General American English. Grammatically, it functions as a nominal noun phrase composed of the classifying adjective agrammatic (modifying the linguistic property of grammar loss) and the abstract mass noun comprehension (denoting the act or capacity of receptive understanding).
In clinical discourse, related morphological variants include the mass noun agrammatism, the agentive noun agrammatic (referring to an individual exhibiting the syndrome, e.g., "the agrammatic patient"), and the adverbial form agrammatically. Orthographically, it is consistently rendered as two unhyphenated words. It functions in clinical documentation and experimental linguistics as both a diagnostic descriptor ("the patient exhibited marked agrammatic comprehension") and an empirical variable within quantitative psycholinguistic paradigms.
4. Detailed Conceptual Explanation
To fully understand agrammatic comprehension, one must dissect the computational mechanisms of human sentence interpretation. Human language understanding operates via two concurrent processing pathways: an interpretive heuristic system based on lexical semantics, real-world pragmatics, and statistical frequency; and an algorithmic syntactic parser that builds structural trees and tracks syntactic dependencies. In non-brain-damaged individuals, these mechanisms cooperate seamlessly. When hearing the sentence, "The apple was eaten by the boy," the listener immediately assigns the agent role to "the boy" and the patient role to "the apple" using both world knowledge (apples cannot eat boys) and syntactic decoding (the passive prepositional phrase "by the boy").
In agrammatic comprehension, this seamless integration is severed. When confronted with semantically reversible sentences, real-world plausibility provides no interpretive crutch. Consider the reversible sentence: "The horse kicked the cow." Either entity is capable of kicking. In canonical English word order (Subject-Verb-Object), agrammatic listeners typically interpret this correctly by relying on a linear, canonical agent-first strategy: assuming that whichever noun appears first must be the thematic agent. However, when the linear order is reversed or disrupted through syntactic movement—such as in passive constructions ("The cow was kicked by the horse") or object-relative clauses ("It was the cow that the horse kicked")—the heuristic strategy fails catastrophically.
Under these non-canonical conditions, agrammatic individuals often exhibit performance at chance levels (roughly 50% accuracy on binary forced-choice sentence-picture matching tasks). This chance performance indicates that they are not simply misinterpreting the sentence systematically; rather, they understand that an action occurred and recognize the participants involved, but they are unable to compute which participant occupies which thematic role. The grammatical markers that signal syntactic movement—such as passive morphology ("was [verb]-ed by") or relative pronouns ("that," "which")—are either not parsed in real time or fail to trigger the trace-licensing procedures required to bind moved noun phrases back to their base positions.
Furthermore, agrammatic comprehension extends beyond simple word order variations to encompass closed-class functional morphemes, grammatical agreements, and binding dependencies. Function words such as determiners, auxiliary verbs, conjunctions, and prepositions frequently fail to be deployed online during sentence parsing. Consequently, agrammatic comprehension is defined not as an inability to understand words, concepts, or general discourse, but as a circumscribed computational deficit in recovering structural, non-canonical, and syntactically displaced thematic relations.
5. Historical Development
The nineteenth-century aphasiological tradition, pioneered by Paul Broca (1861) and Carl Wernicke (1874), established a strict anatomical and functional dichotomy. Broca’s area (the posterior portion of the left inferior frontal gyrus) was designated as the motor-expressive center for articulate speech, while Wernicke’s area (the posterior superior temporal gyrus) was identified as the sensory-receptive storehouse for auditory word images. Under this classical connectionist framework, Broca’s aphasia was defined strictly as an expressive disorder; clinicians assumed that patients’ comprehension was essentially unimpaired because they could follow basic conversational prompts and execute multi-step bedside commands.
This paradigm began to unravel during the mid-twentieth century with the emergence of generative linguistics, founded by Noam Chomsky, and the birth of cognitive psycholinguistics. Researchers recognized that routine clinical evaluations masked severe comprehension deficits because ordinary conversation is saturated with contextual cues, pragmatic redundancies, and canonical word orders. In 1976, Alfonso Caramazza and Edgar Zurif published a groundbreaking study demonstrating that Broca’s aphasic patients performed flawlessly on active semantically irreversible sentences, yet plummeted to chance on semantically reversible center-embedded object-relative sentences.
During the 1980s and 1990s, the field experienced intense theoretical formalization. Yosef Grodzinsky advanced formal syntactic models arguing that agrammatism represented a structural deficit within the linguistic architecture itself, directly linked to damage in Broca’s area. Simultaneously, alternative cognitive architectures were proposed by Eleanor Saffran, Myrna Schwartz, and David Caplan, who framed the impairment in terms of thematic mapping deficits or depleted computational processing resources. In the twenty-first century, structural and functional neuroimaging—particularly functional magnetic resonance imaging (fMRI), magnetoencephalography (MEG), and lesion-symptom mapping—demonstrated that agrammatic comprehension does not result solely from discrete damage to Broca’s area, but rather emerges from disruption across extended left-hemisphere fronto-temporal networks, notably the ventral and dorsal white matter pathways.
6. Theoretical Foundations
Several major theoretical frameworks attempt to explain the underlying cognitive and neurological mechanisms driving agrammatic comprehension. These theories are broadly divided into representational-deficit models, processing-resource models, and mapping models.
Representational-deficit accounts argue that focal brain injury destroys or permanently disables specific syntactic operations or representations. The most prominent example is Yosef Grodzinsky’s Trace Deletion Hypothesis (TDH). Rooted in Government and Binding theory, the TDH posits that individuals with agrammatism are incapable of representing syntactic "traces"—the empty phonological categories left behind when a constituent moves from its deep-structure position to a surface-structure position (such as in wh-questions, relative clauses, and passives). Deprived of traces, the moved noun phrase cannot receive its thematic role syntactically. The patient falls back on a default non-linguistic strategy (assigning the role of "Agent" to the first linear noun), which works for active sentences but generates a conflict in passives, forcing the patient to guess randomly between two perceived agents and producing the classic 50% chance performance profile.
Conversely, processing-resource and capacity-limitation accounts argue that syntactic knowledge remains completely intact, but the patient’s computational working memory, processing speed, or activation capacity is structurally constrained. Formulated by researchers such as David Caplan, Gloria Waters, and Marcel Just, this perspective posits that parsing non-canonical sentences requires heavy real-time working memory overhead to hold displaced constituents in memory while simultaneously constructing the remainder of the syntactic frame. If neural damage reduces processing capacity or slows lexical-syntactic access, the representation of the displaced constituent decays before the parser can integrate it, resulting in parsing breakdowns despite an intact underlying grammatical grammar.
A third framework, the Mapping Hypothesis proposed by Eleanor Saffran, Myrna Schwartz, and Oscar Marin, asserts that agrammatic comprehension represents an inability to map between syntactic positions and semantic thematic roles. According to this view, patients can often build an abstract syntactic parse tree and recognize individual lexical meanings, but the computational bridge that translates structural hierarchies (e.g., subject, direct object) into semantic roles (e.g., agent, patient, theme) is damaged. These competing frameworks continue to stimulate debate regarding whether human language breakdown reflects loss of modular syntactic knowledge or systemic constraints on cognitive implementation.
7. Key Components, Types & Dimensions
Agrammatic comprehension is multifaceted, displaying variations across linguistic structures, diagnostic subtypes, and task demands. The primary dimensions and behavioral components include:
- Canonical vs. Non-Canonical Word Order Discrepancy: The central diagnostic hallmark; patients succeed on canonical Subject-Verb-Object (SVO) structures but fail systematically when syntactic operations invert this sequence to Object-Verb-Subject (OVS).
- Semantic Reversibility Vulnerability: Breakdown occurs specifically when both nouns are animate and capable of performing the action (e.g., "The boy kissed the girl"), disappearing when real-world knowledge makes the thematic assignment irreversible (e.g., "The apple was eaten by the girl").
- Movement-Derived Structures: Marked difficulty processing structures requiring syntactic movement, including:
- Passive Voice Sentences: "The cat was chased by the dog."
- Object-Relative Clauses: "The man whom the woman pushed was tall."
- Object-Cleft Sentences: "It was the zebra that the lion attacked."
- Object Wh-Questions: "Which boy did the girl push?"
- Closed-Class and Functional Morphological Deficits: Inability to exploit grammatical morphemes (prepositions, auxiliaries, tense markings, determiners) to resolve ambiguity during continuous auditory sentence processing.
- Syntactic Complexity Hierarchy: Deficits scale systematically with linguistic complexity, such that center-embedded clauses elicit more errors than right-branching clauses, and double-movement structures yield total incomprehension.
- Subtypes and Associated Syndromes: While historically tied to Broca’s aphasia, agrammatic comprehension is also observed in non-fluent primary progressive aphasia (nfvPPA), specific developmental language disorders, and select cases of subcortical stroke.
8. Examples & Illustrative Cases
To grasp how agrammatic comprehension manifests in clinical reality, consider the following illustrative cases and behavioral profiles.
Case Illustration 1: Stroke-Induced Agrammatic Comprehension
A 58-year-old former civil engineer, Patient M.R., sustained an ischemic stroke resulting in a dense lesion within the left frontal operculum (Brodmann areas 44 and 45) and underlying white matter. In spontaneous conversation, M.R. speaks in halting, telegraphic utterances consisting almost entirely of bare nouns and uninflected verbs ("Wife… hospital… morning… driving"). In ordinary clinic interactions, he appears completely alert and follows complex contextual conversation without hesitation. However, formal neurolinguistic assessment using an audited sentence-picture matching test reveals striking disparities. When presented with the sentence "The elephant pushes the giraffe" alongside two pictures—one showing an elephant pushing a giraffe and the other showing a giraffe pushing an elephant—M.R. scores 100% across 20 trials. Yet, when presented with "The giraffe is pushed by the elephant," M.R.’s performance drops precisely to 52% (11/20 correct). Qualitative observation reveals that M.R. identifies both animals instantly but agonizes over who is performing the push, repeatedly pointing back and forth between the two cards before guessing.
Case Illustration 2: Non-Fluent Primary Progressive Aphasia
A 67-year-old retired school principal, Patient H.K., presents with an insidious, progressive deterioration in speech fluency over three years. Volumetric MRI reveals marked asymmetric atrophy of the left inferior frontal gyrus and anterior insula, leading to a diagnosis of non-fluent/agrammatic primary progressive aphasia (nfvPPA). While single-word comprehension remains pristine, formal syntactic testing demonstrates severe agrammatic comprehension. When presented with an object-cleft construction—"It was the doctor that the nurse phoned"—H.K. consistently selects the picture depicting the doctor making the phone call. Her syntactic parser fails to register the object-cleft marker "It was… that," and she falls back on an agent-first strategy, assigning the role of caller to the doctor simply because "doctor" appeared first in the sentence.
9. Measurement & Assessment
Accurate clinical identification of agrammatic comprehension requires specialized psycholinguistic instruments designed to strip away non-syntactic cues, such as pragmatic context, semantic plausibility, and lexical predictability. Standard bedside aphasia batteries often overlook this deficit because their comprehension subtests frequently evaluate single-word recognition or rely on semantically irreversible commands (e.g., "Point to your nose," "Put the pencil on the paper").
The gold-standard assessment methodology employs Sentence-Picture Matching (SPM) tasks using semantically reversible sentence pairs. In these paradigms, an examiner reads a target sentence or plays a recorded acoustic stimulus, and the patient must select the matching visual scene from a multi-picture array containing the correct target, a thematic role reversal distractor (where agent and patient are swapped), a lexical distractor, and an unrelated foil. Tests such as the Comprehensive Aphasia Test (CAT), the Boston Diagnostic Aphasia Examination (BDAE-3), and the psycholinguistically rigorous Northwestern Assessment of Verbs and Sentences (NAVS), particularly its Sentence Comprehension Test (SCT) subtest, are specifically structured to contrast canonical sentences (active, subject-relative) against non-canonical structures (passive, object-relative, object-cleft).
Beyond picture matching, advanced neurocognitive assessment utilizes Object Manipulation Tasks (e.g., the "acting-out" paradigm), where patients manipulate miniature animal or human figurines in response to spoken commands (e.g., "Make the tiger chase the lion"). This approach prevents patients from relying on visual elimination strategies. In research and specialized clinical centers, eye-tracking while listening (the visual world paradigm) and event-related potentials (ERPs) are employed to monitor syntactic parsing millisecond by millisecond. In neurotypical listeners, syntactic violations or non-canonical dependencies trigger characteristic neurophysiological signatures, such as the early left anterior negativity (ELAN) and the P600 component; patients with agrammatic comprehension routinely exhibit absent, delayed, or degraded P600 waves, demonstrating real-time processing breakdowns during structural reanalysis.
10. Applications & Practical Significance
The practical and clinical significance of identifying agrammatic comprehension is immense, carrying profound implications for diagnosis, targeted neurorehabilitation, and day-to-day communicative interactions.
In speech-language pathology, identifying agrammatic comprehension prevents erroneous clinical classifications. When clinicians mistakenly presume that a non-fluent patient possesses "intact comprehension," family members and healthcare providers are often advised to speak naturally, without modification. This leads to profound miscommunications when patients receive non-canonical directions such as "The blue pill must be taken after the white pill" or "The doctor who was paged by the nurse will see you," which the patient may process backwards. Educating families to use direct, active-voice, canonical phrasing (e.g., "First take the white pill. Then take the blue pill") instantly improves communication and treatment compliance.
In restorative aphasia rehabilitation, this construct forms the theoretical basis for Treatment of Underlying Forms (TUF), an evidence-based intervention developed by Cynthia Thompson and colleagues. TUF directly targets syntactic comprehension and production deficits by training individuals on complex, non-canonical sentence structures derived through syntactic movement (such as object-relative clauses and clefts). Based on the Complexity Account of Treatment Efficacy (CATE), training patients on computationally complex, non-canonical structures reliably generalizes downward to simpler, untrained structures (such as passives and active sentences), whereas training simple structures produces no upward generalization. Consequently, understanding the formal architecture of agrammatic comprehension provides clinicians with the blueprint for designing interventions that stimulate structural neuroplasticity.
11. Research & Empirical Evidence
Decades of empirical research have systematically scrutinized the cognitive mechanisms, lesion correlates, and psycholinguistic validity of agrammatic comprehension. The watershed work of Alfonso Caramazza and Edgar Zurif (1976) first proved experimentally that Broca’s aphasics exhibited near-perfect comprehension on active reversible sentences ("The boy pushes the girl" ~90% accuracy) alongside chance-level performance on object-relative sentences ("The boy that the girl pushes is tall" ~50-60% accuracy). This finding was replicated across dozens of languages, solidifying the empirical reality of the phenomenon.
Neuroimaging and lesion-symptom mapping studies have refined our understanding of its anatomical substrates. While early structural lesion studies tied agrammatic comprehension strictly to damage in Broca’s area (Brodmann areas 44/45), modern voxel-based lesion-symptom mapping (VLSM) studies by Nina Dronkers and colleagues (2004) revealed that severe syntactic comprehension deficits correlate far more strongly with lesions in the left anterior superior temporal gyrus, the temporoparietal junction, and the underlying dorsal white matter fasciculi—specifically the arcuate fasciculus and the superior longitudinal fasciculus. These tracts connect temporal storage areas to frontal parsing machinery, demonstrating that agrammatic comprehension represents a network-level disconnection rather than a localized loss within a single cortical module.
Furthermore, cross-linguistic empirical studies conducted by Elizabeth Bates, Brian MacWhinney, and colleagues within the Competition Model framework provided critical evidence regarding how surface cues affect syntactic parsing. They discovered that while English-speaking agrammatic patients rely heavily on linear word order (agent-first strategy), agrammatic patients speaking richly inflected languages like Italian, German, or Hebrew exploit morphological inflections (case markings, subject-verb agreement) to decipher thematic roles, demonstrating that agrammatic comprehension manifests differently depending on the typological structural cues prioritized by a patient’s native tongue.
12. Cultural & Cross-Cultural Considerations
Cross-linguistic and cross-cultural psycholinguistic research has fundamentally reshaped theoretical models of agrammatism. Early conceptualizations were heavily biased by studies of English, an impoverished language morphologically that relies strictly on fixed SVO word order. In English, when syntactic movement occurs, the parser must rely entirely on abstract hierarchy and closed-class function words.
When aphasiologists examined morphologically rich languages, they observed marked variations in how agrammatic comprehension presents:
- Case-Marking Languages (e.g., German, Russian, Hungarian): In German, grammatical case is explicitly marked on determiners (e.g., der [nominative masculine] vs. den [accusative masculine]). Agrammatic German speakers frequently show preserved ability to decode non-canonical OVS sentences ("Den Hund schlägt der Mann" — The man hits the dog) if the nominative/accusative case markers are phonologically distinct, but fail when case forms are ambiguous.
- Agglutinative and Highly Inflected Languages (e.g., Turkish, Japanese): In Japanese, postpositional particles (such as -ga for nominative and -o for accusative) dictate thematic roles regardless of linear scrambling. Agrammatic Japanese speakers exhibit distinct comprehension profiles, often maintaining understanding as long as the grammatical particles are correctly perceived, but failing when particles are omitted or acoustically masked.
- Tone and Isolating Languages (e.g., Mandarin Chinese): Mandarin lacks inflectional morphology and relies on word order, aspect markers, and pragmatic context. Agrammatic Chinese patients show difficulty with specialized non-canonical constructions, such as the bƮ (disposal) and bèi (passive) structures, highlighting that syntactic comprehension deficits persist across languages regardless of morphological typology.
These cross-cultural findings prove that agrammatic comprehension is not an artifact of English sentence structure, but rather a universal neurobiological vulnerability that exposes the distinct computational pressure points of whichever language the speaker has internalized.
13. Criticisms, Debates & Limitations
Despite its central role in aphasiology, the construct of agrammatic comprehension has faced significant criticism, theoretical debates, and methodological challenges.
The primary controversy surrounds syndromic validity and individual variation. Classical aphasiology treated "Broca’s aphasics with agrammatism" as a homogeneous clinical group. However, extensive single-case studies by David Caplan, Eleanor Saffran, and Alfonso Caramazza demonstrated radical individual variability. Some patients who produce severely agrammatic, telegraphic speech show intact comprehension of complex non-canonical sentences, while other patients with fluent speech production exhibit classic agrammatic comprehension profiles. This dissociation led Caramazza to argue against treating agrammatism as a unified, biologically valid syndrome, suggesting instead that it represents an arbitrary cluster of distinct cognitive deficits that happen to co-occur following large middle cerebral artery strokes.
A second major debate concerns the nature of the chance-level performance observed in experiments. Proponents of representational accounts, such as Grodzinsky, interpret 50% performance on passive sentences as formal proof of missing syntactic traces and structural guessing. Critics, such as Caplan and Waters, argue that 50% group accuracy is often a statistical artifact produced by averaging together patients who consistently pass, patients who consistently fail, and patients who vacillate due to fluctuating attention. They contend that resource-reduction models—which posit intermittent processing capacity failures during high cognitive load—offer a far more biologically plausible explanation than absolute, categorical trace deletion.
Finally, critics emphasize task-demand artifacts. Traditional Sentence-Picture Matching forces patients to process complex pictorial scenes while parsing spoken language, introducing visual search, executive selection, and working memory demands that are separate from linguistic competence. When tested using online, implicit tasks that minimize executive overhead (such as auditory word-monitoring or self-paced listening), agrammatic patients often demonstrate sensitivity to syntactic violations that they fail to decipher during offline, conscious picture-selection tasks.
14. Related Terms & Distinctions
To ensure precise diagnostic and conceptual application, agrammatic comprehension must be differentiated from related neurolinguistic and cognitive phenomena:
- Expressive Agrammatism: The production deficit characterized by telegraphic speech, omitted functional morphemes, and simplified phrase structure. While expressive and receptive agrammatism frequently co-occur, they dissociate double-directionally in neurogenic populations.
- Paragrammatism: Typically observed in Wernicke’s or fluent aphasias, paragrammatism involves fluent speech containing inappropriate juxtapositions, grammatical confusions, and morphological substitutions, rather than the systematic omissions and structural reductions seen in agrammatism.
- Lexical-Semantic Comprehension Deficit: A breakdown in retrieving the core semantic concepts of individual words (observed classically in Wernicke’s aphasia and semantic dementia). Patients with agrammatic comprehension have intact lexical-semantic knowledge; they know what an "elephant" and a "giraffe" are, but fail to parse how they relate structurally.
- Working Memory Impairment (Non-Linguistic): A general deficit in short-term cognitive storage and manipulation. Although verbal working memory constraints can exacerbate agrammatic comprehension, the latter is distinguished by its specific vulnerability to syntactic movement and non-canonical thematic assignment.
- Auditory Agnosia / Pure Word Deafness: A low-level acoustic-phonetic processing deficit where patients cannot distinguish speech sounds despite intact hearing. Agrammatic comprehension operates entirely at the higher linguistic-structural level, with phonological processing intact.
15. Summary / Key Takeaways
Agrammatic comprehension represents a selective failure in algorithmic syntactic parsing, leaving patients unable to decipher semantically reversible sentences whose thematic assignments depend strictly on non-canonical word order, syntactic movement, and functional morphology. Rooted historically in the re-evaluation of Broca’s aphasia during the cognitive revolution, it transformed aphasiology from simplistic motor-versus-sensory categorizations into a nuanced, computationally grounded science. Whether interpreted through the lens of representational loss (such as trace deletion), processing resource exhaustion, or thematic mapping failure, the condition highlights the distinct neural architecture dedicated to structural language computation. Accurate diagnosis requires psycholinguistically balanced assessments like sentence-picture matching with reversible passive and relative constructions. Finally, targeted therapeutic protocols such as the Treatment of Underlying Forms leverage this syntactic architecture to drive meaningful linguistic recovery, underscoring the vital bridge between theoretical neurolinguistics and clinical neurorehabilitation.
References
- Caramazza, A., & Zurif, E. B. (1976). Dissociation of algorithmic and heuristic processes in language comprehension: Evidence from aphasia. Brain and Language, 3(4), 572–582. https://doi.org/10.1016/0093-934X(76)90048-1
- Caplan, D., & Waters, G. S. (1999). Verbal working memory and sentence comprehension. Behavioral and Brain Sciences, 22(1), 77–94. https://doi.org/10.1017/S0140525X99001788
- Dronkers, N. F., Wilkins, D. P., Van Valin, R. D., Redfern, B. B., & Jaeger, J. J. (2004). Lesion analysis of the brain areas involved in language comprehension. Cognition, 92(1–2), 145–177. https://doi.org/10.1016/j.cognition.2003.11.002
- Grodzinsky, Y. (2000). The neurology of syntax: Language use without Broca’s area. Behavioral and Brain Sciences, 23(1), 1–21. https://doi.org/10.1017/S0140525X00002399
- Thompson, C. K., & Shapiro, L. P. (2007). Complexity in treatment of syntactic deficits in aphasia. American Journal of Speech-Language Pathology, 16(1), 30–42. https://doi.org/10.1044/1058-0360(2007/005)