Ken Ramshøj Christensen

Negative and affirmative sentences increase activation in different areas in the brain

Abstract Though negation is unique and central to human language, it has so far received little attention in cognitive neuroscience. The goal of the present study was to investigate the contrast between affirmative and negative sentences using fMRI, focusing on two central aspects, namely, (1) a semantic difference: affirmation is upward entailing, whereas negation is downward entailing; (2) a syntactic difference: negation involves more syntactic structure than affirmation. The behavioural data showed that negation significantly increased response times (but not the level of performance), even when negation was only in the preceding context to the response condition. The imaging results showed increased activation in the left premotor cortex from negation, compatible with rule-governed memory processing, and increased activation in the right supramarginal gyrus from affirmation, compatible with semantic processing. Finally, affirmation showed “default mode” activation in the cingulate cortex.

The locative alternation: distinguishing linguistic processing cost from error signals in Broca's region.

The left inferior frontal gyrus (LIFG) is known to be involved in the processing of syntactic complexity, such as word order variation. It is also known to be involved in semantic interpretation in studies of various types of semantic and pragmatic anomalies. Across neuroimaging studies of language processing, two main approaches can be found, one that contrasts anomalous and well-formed words or sentences in order to yield an error response and one that contrasts two well-formed syntactic structures differing in complexity, investigating effects of increased integration costs. The present fMRI study aimed at disentangling the error signal from the processing cost signal in LIFG. To do so, we examined the so-called Locative Alternation, which involves the contrast between the Content-Locative construction, e.g. He sprays paint on the wall, and the Container-Locative construction, e.g. He sprays the wall with paint, which have been argued to differ in processing. By including asymmetric verbs, e.g. He blocks the road with rocks vs. *He blocks rocks on the road, we were able to study the contrast between well formed and anomalous constructions. Participants performed an acceptability judgment task during fMRI. The results showed that increased syntactic integration costs yielded both increased response time as well as LIFG activation. Anomalous sentences yielded low acceptability rating but no increase in response time, yet they also evoked increased LIFG activation. Thus, the processing cost and the error signal were found to be functionally independent, but spatially overlapping in the brain.

Escape from the island: grammaticality and (reduced) acceptability of wh-island violations in Danish.

In the syntax literature, it is commonly assumed that a constraint on linguistic competence blocks extraction of wh-expressions (e.g. what or which book) from embedded questions, referred to as wh-islands. Furthermore, it is assumed that there is an argument/adjunct asymmetry in extraction from wh-islands. We report results from two acceptability judgment experiments on long and short wh-movement and wh-extraction from wh-islands in Danish. The results revealed four main findings: (1) No adjunct/argument asymmetry in extraction from wh-islands. (2) Long adjunct wh-movement is less acceptable than long argument wh-movement, and this difference is attributable to matrix verb compatibility and factivity, not D-linking. (3) Long movement reduces acceptability, but is more acceptable than island violations. (4) Training effects reveal that island violations, though degraded, are grammatical in Danish. Since the standard assumptions cannot account for the range of results, we argue in favor of a processing account referring to locality (processing domains) and working memory.

Interfaces, syntactic movement, and neural activation: A new perspective on the implementation of language in the brain

Abstract Studies of language deficits as well as neuroimaging studies indicate that syntactic processing of displaced constituents is implemented in the brain as a distributed cortical network of modules. The data from the present fMRI study on two types of syntactic movement in Danish offers further support for such a distributed syntactic network. These results, together with the results from a number of other fMRI studies in the literature, form the basis for the Domain Hypothesis according to which differential activation in the subcomponents of the cortical network reflects computation of different syntactic domains—the interface levels between syntax, semantics, and pragmatics. The activation patters result from the interaction between movement and target domain, not (non-) canonicity or working memory per se. Specifically, movement to the CP-domain activates areas including Brocas area and the temporal–occipital–parietal junction, whereas movement to the IP-domain does not, but activates the left anterior temporal cortex.

On the nature of escapable relative islands

It is generally assumed that universal island constraints block extraction from relative clauses. However, it is well-known that such extractions can be acceptable in the Scandinavian languages. Kush & Lindahl (2011) argue that the acceptability in Swedish is illusory; relative clauses that allow extraction have a different structure (small clause structure) from those that block extraction (true relatives, CPs). We present data from an acceptability survey of relative clause extraction in Danish. In the survey, extraction significantly decreased acceptability but we found no statistically significant effect of the ability of the verb to take a small-clause complement. We also found no difference between som ‘that/who/which’ and der ‘that/who/which’, both of which can head a relative clause while only som can head a small clause. We argue that our results do not warrant the stipulation of a structural contrast between acceptable and unacceptable extractions, and that variation in acceptability stems from processing.

Structure before Meaning: Sentence Processing, Plausibility, and Subcategorization

Natural language processing is a fast and automatized process. A crucial part of this process is parsing, the online incremental construction of a syntactic structure. The aim of this study was to test whether a wh-filler extracted from an embedded clause is initially attached as the object of the matrix verb with subsequent reanalysis, and if so, whether the plausibility of such an attachment has an effect on reaction time. Finally, we wanted to examine whether subcategorization plays a role. We used a method called G-Maze to measure response time in a self-paced reading design. The experiments confirmed that there is early attachment of fillers to the matrix verb. When this attachment is implausible, the off-line acceptability of the whole sentence is significantly reduced. The on-line results showed that G-Maze was highly suited for this type of experiment. In accordance with our predictions, the results suggest that the parser ignores (or has no access to information about) implausibility and attaches fillers as soon as possible to the matrix verb. However, the results also show that the parser uses the subcategorization frame of the matrix verb. In short, the parser ignores semantic information and allows implausible attachments but adheres to information about which type of object a verb can take, ensuring that the parser does not make impossible attachments. We argue that the evidence supports a syntactic parser informed by syntactic cues, rather than one guided by semantic cues or one that is blind, or completely autonomous.

The Infinitive Marker across Scandinavian

In this paper I argue that the base-position of the infinitive marker in the Scandinavian languages and English share a common origin site. It is inserted as the top-most head in the VP-domain. The cross-linguistic variation in the syntactic distribution of the infinitive marker can be accounted for by assuming that it undergoes head movement. This movement is optional in Danish, English, Norwegian, and Early Modern Danish and is not feature-driven. In Faroese, Icelandic, and Swedish, on the other hand, it is triggered by φ-feature checking on Fino. In Icelandic and Swedish these φ-features are strong and induce obligatory vo→Fino movement, whereas they are weak in Faroese and do not induce vo→Fino movement.

(Non-)Arguments in Long-Distance Extractions.

Previous research has shown that in fully grammatical sentences, response time increases and acceptability decreases when the filler in a long-distance extraction is incompatible with the matrix verb. This effect could potentially be due to a difference between argument and adjunct extraction. In this paper we investigate the effect of long extraction of arguments and adjuncts where incompatibility is kept constant. Based on the results from two offline surveys and an online experiment, we argue that the argument/adjunct asymmetry in terms of acceptability is due to differences in processing difficulty, but that both types of extraction involve the same intermediate attachment sites in the online processing.

A Verbal Illusion: Now in Three Languages

The so-called depth charge sentences (e.g., no head injury is too trivial to be ignored) were investigated in a comprehension experiment measuring both whether participants understood the stimuli and how certain they were of their interpretation. The experiment revealed that three factors influence the difficulty of depth charge type sentences: the number of negations, the plausibility of the relation between the subject and the verb, and finally the logic of the relation between the adjective and the verb. When a sentence is maximally complex (i.e., when there are multiple negations, the relation between subject and verb is implausible, and the relation between adjective and verb is illogical) participants misunderstood the sentence, but were at the same time certain of their answers. The experiment supports the idea that depth charge sentences create a verbal illusion—the sentences mean one thing, but people systematically understand them to mean the opposite.

CP-recursion in Danish: A cP/CP-analysis

Abstract Based on data from extraction, embedded V2, and complementizer stacking, this paper proposes a cP/CP-analysis of CP-recursion in Danish. Because extraction can be shown to be possible from relative clauses, wh-islands, and adverbial clauses, and given that long extraction is successive-cyclic, an extra specifier position has to be available as an escape hatch. Consequently, such extractions require a CP-recursion analysis, as has been argued for embedded V2 and for complementizer stacking. Given that CP-recursion in embedded V2 clauses does not allow extraction, whereas other types of CP-recursion do, we suggest that embedded V2 is fundamentally different, in that main clause V2 and embedded V2 involve a CP (“big CP”), whereas all other clausal projections above IP are instances of cP (“little cP”). The topmost “little” c° has an occurrence feature that enables extraction but bars spell-out of its specifier.