(traduzir, editar e organizar)
O componente de ERP, o N400, é uma das assinaturas fisiológicas mais robustamente replicadas, e também o componente mais amplamente estudado em pesquisas linguísticas com metodologia ERP. Apresenta-se como um pico de amplitude negativas (geralmente plotada para cima) que emerge por volta de 400ms após o início de apresentação do estímulo, que geralmente é uma palavra, isolada ou em contexto de sentença, falada ou escrita.
O N400 foi encontrado “por acidente” em 1978, quando os psicólogos Kutas e Hillyard estavam realizando um experimento do tipo oddball, mas em vez de usar estímulos visuais simples (como era de costume com esse paradigma experimental) , eles decidiram tentar usar estímulos linguísticos, a saber, sentenças. Eles tinham a expectativa de que finais imprevisíveis de frases (por exemplo, He spread the warm bread with socks; ‘Ele espalhou no pão quente ‘meias’) produziriam um efeito P300, que é a resposta neurofisiológica usual para esse paradigma. Para sua surpresa, eles encontraram um componente mais tardio, um pico negativo cerca de 400 ms após o início da apresentação da palavra, o qual ficou conhecido como o N400. O ‘efeito de N400’ ficou caracterizada principalmente por uma amplitude alta (negativa) de N400 associada de forma geral à dificuldade de integrar palavras semanticamente implausíveis em sentenças(Figura 1).
The cognitive specificity of these symbolic representations, and the possibility of testing hypotheses regarding the serial nature of language processing owing to the quintessential online nature of ERP measures quickly sparked researchers’ interest. Already within the first decade of its discovery, scientists confirmed the robustness of the N400 component in auditory and visual modalities (KUTAS, M. VAN PETTEN, 1988; HOLCOMB & NEVILLE, 1990; BROWN, C. HAGOORT, 1993), with a range of languages from different ‘language families’, for instance English as well as Finnish (RELANDER et al., 2009), with spoken or signed language (NEVILLE et al., 1997; KUTAS & FEDERMEIER, 2011), and at a wide range of presentation durations or SOAs4 , ranging from 40 to 250ms (KUTAS, M. VAN PETTEN, 1988; BROWN, C. HAGOORT, 1993; ROSSELL et al., 2003; GREWE et al., 2007; LAU et al., 2008; KUTAS & FEDERMEIER, 2011). Also from the early 90s, studies showed that by manipulating strictly lexical features, such as word frequency, word repetition, phonotactic probability, or morphological priming (OSTERHOUT et al., 1997; PYLKKÄNEN, 1997; MUNTE et al., 1999; KUTAS & FEDERMEIER, 2000a; PYLKKÄNEN et al., 2000; STRINGFELLOW & MARANTZ, 2002; DOM et al., 2004; MARANTZ, 2005a; FRANÇA, A.I., LEMLE, M. GESUALDI, A.R, CAGY, M. INFANTOSI, 2008; LAU et al., 2008; SOTO, 2010; GARCIA et al., 2012), in word stimuli, presented in isolation or in priming pairs5 also rendered N400 effects, this time with lower peak amplitudes associated to facilitated access (SOTO, 2014).
Embora o componente tenha sido encontrado desde então com uma variedade de tipos estímulos (não só linguísticos), por exemplo, com imagens (para uma revisão cf. KUTAS e FEDERMEIER, 2011), uma violação da expectativa da melodia não resultará em um efeito N400 (KUTAS e VAN PETTEN, 1988). Portanto, Kutas e Federmeier (2011), concluem que o N400 reflete acesso ao conhecimento semântico na memória de longo prazo, de natureza multimodal. Kutas, for instance, refers to the N400 as reflecting access to a “knowledge base, known as ‘semantic memory’, often in response to a linguistic cue in the form of a spoken, written or signed word” (KUTAS & FEDERMEIER, 2000b: 436). The notion of linguistic content as a ‘cue’ to rich conceptual structures is typical of theories that are semantic-centered; formal theories, on the other hand, emphasize the importance of language structure in constructing and organizing meaning, for example, in the way entities and events are marked by lexical categories (marry-V vs. marriage-N vs. marital-A), in the way events are structured by thematic grids (number of arguments and syntactic positions projected by the verb and their semantic interpretation: e.g. x-agent (marry (y-patient))) and case marking (e.g. xNOM(invite(y-ACC))), or in the way duration of events are interpreted through aspect and object noun type (e.g. eating the cake in ten minutes vs. eating cake all day), etc. These examples show how meaning can be mediated strongly by language structure (SOTO, 2014). This does not mean that conceptual knowledge does not exist without language. On the contrary, it does. But the question is whether the architecture of language equips humans with a special way to make sense of the world and how and when we tap into the world knowledge through syntactic categories (aspect, tense, numerosity, etc.) and their combinations (SOTO, 2014).
Mas se pensarmos no processamento linguístico, e a natureza incremental desses processos, junto com o fato de componentes podem refletir atividades somadas (i.e. eletrodos captam atividade que pode ser resultante de engajamentos neuronais que se diferenciam ligeiramente em fonte geradora ou momento de ocorrência), o N400 pode refletir ora o resultados compilado de múltiplas computações cognitivas subjacentes ou refletir apenas uma computação em uma série sequencial de fases de processamento (KUTAS and FEDERMEIER 2011; LAU, PHILLIPS and POEPPEL 2008).
The first N400 studies in the 80s replicated results found by Kutas and Hillyard (1980) who presented subjects with sentences such as I take my coffee with cream and dog and found higher more negative amplitudes at 400ms after word-onset (of dog) in comparison to the same sentence with the semantically compatible ending ...and sugar, thus, neurophysiologically marking semantically implausible endings to sentences. In the many studies that followed in the 80s and 90s (KUTAS, M. VAN PETTEN, 1988; OSTERHOUT et al., 1997; BERKUM et al., 1998; FEDERMEIER & KUTAS, 1999; KUTAS & FEDERMEIER, 2000b), scientists discovered that the N400 amplitude varies according to the degree of semantic (im)plausibility (e.g. higher amplitudes for Joe ate a shoe than for the less plausible, but slightly possible Joe ate a Coca-cola), directly linking higher voltage values with increased cognitive effort to integrate ill-fitting lexical items in preceding contexts. Given that the effect had been found in the context of complement selection by verbs and prepositions (both theta-role assigning predicates), linguists quickly correlated the N400 component to theta-role attribution, as it marks the moment of verb-complement merging as a result of the projection of the verb’s thetagrid (CHOMSKY, 1993). Results obtained by França et al. (2004) showed that N400 is also modulated by long distance dependencies between verbs and their complements. For example, in Wh-questions Which songs will he sing vs. 11?, França, 2004), pronominal referencing (He took the knife and he’s going to sharpen/cut it, noting that, in Portuguese, pronominal references agree in gender, strengthening the violation, França, Soto and Gesualdi, 2012), further confirming the involvement of semantic-syntactic dynamics (SOTO, 2014).
To complicate the picture even more, in the 90s, researchers started investigating the N400 in the context of priming and isolated word presentation. In these experiments, purely lexical variables, such as repetition, frequency and phonotactic probability, seemed to affect the component (HOLCOMB & NEVILLE, 1990; BROWN, C. HAGOORT, 1993; KUTAS, 1993; PYLKKÄNEN et al., 2000; KUTAS & FEDERMEIER, 2000b; LAU et al., 2008), equally correlated with relatively increased or decreased cognitive effort. Hence, hypotheses emerged linking the N400 also to lexical access. In fact, by framing the N400 effect as mainly a question of predictability, rather than implausibility, relative amplitude increase or decrease to words in sentences could just as easily be attributed to either facilitation due to high predictability or strong mismatch effects when the upcoming word strongly differs from the predicted lexical item. In fact, some studies show that the N400 is indeed mainly graded by facilitation effects, enhancing activation of lexical candidates according to predictability by the context, and less so by restricting the set of possibly activated candidates, for example by constraining the sentence context, either semantically or syntactically (Federmeier et al., 2007, GASTON et al., 2019).
O efeito N400 é modulado por variáveis linguísticas. Um exemplo é priming morfológico (ex. jardim-jardinheiro), que ocorre dissociado de priming ortográfico ou semântico com priming coberto ou priming comum (FRANÇA et al.,2008; GOMES, FRANÇA, 2013, SOTO, 2017). Outros exemplos são efeitos de complexidade sintática, como com perguntas Qu- ou com correferência pronominal (e.g. ‘Ele pegou a faca e vai afiar/cortá-la’, é importante notar que, em português, as referências pronominais concordam em gênero, reforçando a violação semântica de ‘cortá-la (i.e. a faca), FRANÇA, SOTO, GESUALDI, 2012). Dadas as evidências de envolvimento sintático nos processos subjacentes ao N400, geralmente associados a processos semânticos (FEDERMEIER et al., 2013; FRANKLIN et al., 2007; KUTAS e FEDERMEIER 2011; LAU et al. ., 2009), o fenômeno deve ser considerado, na verdade, na interface sintaxe-semântica. Nesse sentido, a resposta a uma palavra-alvo (como chá vs. chalé na frase 'João bebeu o chá vs. chalé', FRANÇA et al., 2002) reflete o momento em que o verbo (ou qualquer outro predicador projetando um complemento, como uma preposição) atribui um papel teta, marcado morfossintaticamente por caso, licenciando (ou deixando de fazê-lo, no caso de violação) seu(s) argumento(s).
Nevertheless, some studies do attest to the influence
of different levels of context. For example, with sentences
such as
They wanted to make the hotel look more like a
tropical resort. So along the driveway they planted rows
of ... palms vs. pines vs. tulips,
Federmeier and Kutas
(1999) studied the effect of sentential context on endings
varying in improbability. They found a modulation of
the effect in which the word
tulips
shows the highest
amplitude due to incompatibility with preceding sentential
context (characterized by
tropical resort
and
driveway
);
whereas pines, equally improbable, showed intermediate
amplitudes, according to the authors due to its semantic
relatedness to
palms
(i.e. also a tree), the preferred ending.
This would illustrate what is known as surface reading:
rather than going deep into the hierarchical structure
embedded within the sentence, a few key semantic features
taken from individual word roots are interpreted to construct
a sort of surface reading. Other readings may derive from
this but are not thought to bootstrap processing.
Another study by Hagoort et al.
(2004), manipulated
conceptual knowledge vs. world knowledge, where
conceptual violations
(Dutch trains are sour)
yielded
similar N400 effects as world knowledge violations
(Dutch trains are white,
when in reality in Holland all
trains are yellow). These examples are among many that
indicate that higher level information and surface reading
processing might affect the N400. If this is true, then either
higher level information is available during syntactic
processing, which would prove hypotheses claiming strict
serial and modular processing streams to be incorrect,
or
that such N400 effects reflect post syntactic processes.
In what context does context matter?An ERP study of sentence processingin Brazilian PortugueseEm que contexto o contexto importa? Um estudo de ERP com sentençasem português brasileiroMarije SotoAniela Improta FrançaJuliana Novo GomesUniversidade Federal do Rio de Janeiro – Rio de Janeiro – Rio de Janeiro – Brasil Aline Gesualdi Manhães
Letras de Hoje
, Porto Alegre, v. 59, n. 1, p. 120-130, jan.-mar. 2015
SOTO, M.; FRANÇA, A. I. THE FUNCTIONAL ANALYSIS OF THE N400 COMPONENT: LEXICAL ACCESS, INTEGRATION OR CAN WE HAVE IT BOTH WAYS? Revista LinguíStica, Rio de Janeiro | Volume 16 | Número Especial Comemorativo | p. 521-562 | nov. 2020
Comments