The brain of homo sapiens acquired the ability of reading and writing after the long-standing phylogenetical development. The invention of letters more than 5,000 years ago by our ancestors has realized the communication beyond the limit of time and space, and accordingly all the living creatures finally entered into the histological age of evolution. Thus, the cerebral mechanism of reading and writing might be one of the ultimate goals of the evolution of brain.
Scientific approach to the study of cerebral mechanism of reading and writing was initiated by the clinical observations of the patients who had become unable to read and write due to focal brain damage. Through these classical studies, the angular gyrus theory was proposed in the western world to explain the brain function underlying the ability of reading and writing. Although the angular gyrus theory was successful in explaining the anatomo-clinical features of alexia (disability in reading) and agraphia (disability in writing) in westerners, there had been found cases of alexia and/or agraphia in Japan whose clinical pictures could never been explained by this simple classical theory.
The enigma was solved when a patient with alexia and agraphia only in kanji with preservation of reading and writing in kana was found to have a focal damage in the left posterior inferior temporal lobe, which was thought to be involved in the process of semantic reading as well as visual evocation of letter form in writing. Upon the clinical observations we proposed the dual processing model of cerebral mechanism in reading and writing. Our recent investigations of PET scan activation studies are now confirming the validity of this dual processing model.

抄録全体を表示

In this overview, we will review the past paths of development of linguistic theories, particularly those in the so-called generative tradition of grammar, and attempt to foresee the natural consequences of these in the near future.
After more than a quarter century of its history, generative grammar has experienced a tremendous shift in its methodology, which could be summarized as the trend from rules to principles. Thus, current linguistic theory is not a theory of grammar rules, but one about constraints on linguistic information that are assumed to be encoded in the human brain system.
Even though there are different conceptions about the nature of linguistic constraints among linguists, many tend to agree that these constraints are quite similar to those found in physical world, e.g., constraints in mechanics. In this respect, linguistic theory can be considered to describe a certain type of law that exists inside one of the smallest universes, viz., our brain.
Consequently, the linguistic theory is shifting toward a more general theory about abstract linguistic objects. It is a theory of how a human being deals with information for conceptualization and communication, though there are some who doubt whether language is sufficiently effective for the latter purpose.

抄録全体を表示

作業記憶 (working memory) は, これまでは主として情報を受動的に保持する機能に着目して論じられていた. 短期記憶 (short-term memory) とは, そうした観点から作業記憶をとらえた概念である. しかし, 保持という視点からだけでは説明できない認知的現象がある. また, 伝統的な意味での短期記憶の容量と, ある種の認知的作業の成績との間に相関がないこともわかっている. そこで本稿では, 情報の保持と処理の両方にわたる計算資源の供給元として作業記憶をとらえ, 操作的な観点からその容量の制約を考えることにより, 統一的な認知モデルにおいて重要な位置を占めるべき作業記憶の理論を探る.
高次の認知活動における作業記憶の役割を考える上で重要なのは, 課題が要求する情報の保持と処理の量が多ければ多いほど, 作業記憶の容量の差がより顕著に現れるという点である. たとえば, 言語理解における困難は作業記憶の容量の制約と相関関係があるが, 埋め込み文や曖昧語の理解において, 理解が困難な場合の方が作業記憶の容量に関する個人差の影響も大きい. 問題解決や空間的思考に関しても同様に, 課題が複雑になるほど作業記憶の容量の差が成績に大きく反映する. 異種の認知過程において作業記憶の容量の制約がこのように同様の仕方で現れていることから, 作業記憶に関してはさまざまな認知領域にわたる共通の説明が望まれる.
著者らが提案している認知の計算アーキテクチャ3CAPSは一種の並列プロダクション・システムであり, そこでは, 情報の保持と処理の両方が活性 (activation) という共通の計算資源に基づく. プロダクション規則が発火するには, 条件部の指定する要素が作業記憶中に保持されている必要があるが, それには, その要素が単に存在するだけではなく, ある程度以上の活性を持たなくてはならない. また, 規則が発火すると出力要素の活性が高まるが, 3CAPSではこの活性化が一挙にではなく少しずつ行なわれる. 作業記憶の容量に関する制約は, この活性の総量の限界から生ずる. すなわち, この限界に近い状態では, 規則の出力要素の活性化が遅くなって処理が阻害されると同時に, 既存の要素の活性が下がって忘却が生ずる. このモデルによって上述のようなさまざまな認知現象が統一的に説明できる.
さらに, 健常者と失語症患者との差, および若者と老人との差を, いずれも知識の差によってではなく, 作業記憶の容量の大小という定量的な概念によって説明できると考えられる. 実際, 適当なポーズを文に挿入することによって失語症患者による文理解の成績が向上するのに対し, 厳しい時間制約の下では健常者も失語症患者と同種の理解の障害を示す. 加齢による認知能力の低下も, 知識や技能の単なる想起に関してはあまり目立たず, 作業記憶に大きな負担を強いるような認知過程においては著しい. このことは, 認知における知識や技能の蓄積の効果と作業記憶の効果とが分離可能であることを示唆するゆえに, 特に興味深い.

抄録全体を表示

This paper introduces a trend of recent AI researches and addresses a view of future AI researches which are relevant to cognitive science. The keywords in recent AI researches are ‘Large’, ‘Fast’, and ‘Move’. These keywords seem to imply that AI are aiming at making a new intelligence which is unique to machine. The author thinks, however, that cognitive science will still play an essential role in artificial intelligence. ‘Heuristics’ is the most important in AI. AI researchers will have to study how to get (or select or create) good heuristics and how to represent them.

抄録全体を表示

This article discusses past developments and future prospects of cognitive engineering in human-computer interaction (HCI). One aim is to shed some light on design problems that may arise as we extend our scope of HCI design from the immediate interface between a single user and the system, to the indirect interface between the system and the task world, and to the organizational interface for cooperative work between people and systems. Also discussed are methodological problems associated with the attempt to build more or less general models of the user in HCI. Lastly, implications of future technological advancements in HCI for a cognitive science view of human communication are discussed in a somewhat futuristic manner.

抄録全体を表示

Roles of subjective occluding surfaces and transparency by the neon effect in the formation of amodal representation of partly-occluded objects were examined in terms of their effects on the solution to the correspondence problem in apparent motion. One of the following horizontally-elongated surfaces was located between the top and the bottom motion targets in the bistable motion display: a textured rectangle (real surface), an opaque illusory surface, or a transparent illusory surface induced by the neon brightness effect. The binocular disparity of the surface was varied. Proportion of transversal motion perception increased when a real or an opaque illusory surface carried a crossed disparity, but not when a transparent illusory surface was used. The results suggest that the formation of amodal representation occurs in the real or the illusory surface cases but not in the transparent surface case, resulting in the difference in affinity between the top and the bottom motion targets in the bistable motion display. Effects of nonlocal information about surfaces on the formation of amodal representation and nonhierachical organization of surface representation have been demonstrated.

抄録全体を表示

What makes it a pleasure to interact with a computer system? In this paper,we discuss some possible factors contributing to a pleasant interaction.
We study the pleasure of interaction under the domain of a two-player game called word tennis, Shiritori in Japanese, conducted through a computer network. Two sets of experiments were carried out. The first set consisted of game sessions between a subject and a computer player. Half of the subjects were informed (correctly) that their opponents were computer players, while the other half were informed (incorrectly) that their opponents were human players. All subjects could quit the game session whenever they chose to do so. The second set was the same as the first one except that the computer player was replaced by a human player.
From the results of both sets of experiments, we reveal that, regardless of whether the actual identity of the opponent is a computer or a human player, subjects who were told to be facing with a human player gave significantly higher pleasure-ratings than those who were told to be facing with a computer player. Furthermore, subjects who found the game pleasant kept playing the game until the end of the pre-determined time limit (one hour). As the word tennis game is dry and monotonous in essence, the exceptionally long game session suggests that the subjects must have had an intrinsic motivation for the game. However, previously recognized intrinsic motivation for games in general fails to explain our observed phenomenon satisfactorily. Hence, we make an attempt to characterize a novel intrinsic motivation, and postulate that the cognition of being involved communication with a human mind is indeed a valid intrinsic motivation in interacting with a computer system.

抄録全体を表示

This article discusses 1) theoretical models of text revision, 2) skills required in revision, and 3) effects of writing support systems, including wordprocessors. Revision is not only a corrective process which comes after writing but a recursive reevaluation of the text's intended meaning and structure, occurring over the whole course of writing. Revision basically consists of detecting problems in a text and diagnosing those problems to discover ways of solving them. Diagnosing calls for more skills than detecting does. A writer's personal definition of revision has a strong influence on the processes and results of his revising; as does his ability to make a strategic use of skills and knowledge. Whereas an experienced writer tries to discover in a text what he really wants to communicate to readers and how to express that, a beginning writer identifies revision with rewording - a step which comes after he has already determined what to say. Word processors were expected to improve students' revision by reducing the physical workload in editing a text, which would encourage students to spend more of their cognitive resources in reevaluating a text's structure and coherence. Studies which examined the effects of using a word processor in comparison with pen and paper writing did not support that expectation. When a writer uses a word processor to write and revise a text, he spends less time in planning than when he uses pen and paper. Attention is focused on small segments, rather than on the whole of the text. In addition, writers find it easier to read and understand a text on paper than one on a computer display. In conclusion, also discussed are educational and technological methods for facilitating and improving revision.

抄録全体を表示