在這些數位學習環境中，歸納式學習與演繹式學習為最常見的兩種教學方式。歸納式學習指的是一種由下而上（bottom-up）的學習方式，學習者透過觀察語言例句，主動發現其中的規律並歸納出文法規則。而演繹式學習則是一種由上而下（top-down）的學習方式，學習者會先被給予明確的文法規則說明，再提供例句與練習。然而，這兩種教學方式各有其優缺點：歸納式學習雖能促進深度理解，但可能增加學習負擔，演繹式學習雖能使學習者能夠直接掌握學習重點，進而提升整體學習效率。然而卻可能降低學習者主動思考的機會。因此，本研究藉著整合歸納式學習與演繹式學習的優勢來發展Inductive and Deductive English Architecture (IDEA)。前者透過讓學習者閱讀電子文章來探索文法規則，而後者則是藉著文法繪本來明確講解文法規則。
Study 1的研究結果顯示，多數學習者皆認為IDEA 1.0有助於學習者能以較直觀的方式理解抽象的文法概念，進而提升其學習動機與投入程度。然而，整體型學習者的負面觀感較少、重視介面的美感與互動且偏好直覺操作與自由探索，而序列型學習者提出較多負面觀感，且對於格式結構與細節較為敏感，且需要清楚的結構指引。
綜上所述，整體型與序列型學習者最終皆能獲得相近的學習成效，但在學習觀感、行為頻率與行為序列仍呈現差異。因此，本研究的貢獻在於透過電子閱讀以及文法繪本來實現歸納式學習與演繹式學習之整合，能夠讓不同認知風格的學習者選擇自己喜好的方式來學習英語文法，同時兼顧他們的學習成效以及任務成效，所以數位學習的設計者在未來也可以使用此整合型的學習方式在不同學科中，以在這些學科兼顧不同認知風格學習者的需求和喜好。;With the rapid development of globalization and digital technologies， English has become one of the most important languages for international communication and knowledge exchange. Among various aspects of English learning， grammar serves as a foundation of language acquisition， enabling learners to produce correct sentences. However， non-native English learners may think that grammatical rules are often abstract and complex， causing difficulties in comprehension and reducing learning motivation. To address these challenges， digital learning approaches have been incorporated into English grammar learning for recent years.
Among them， inductive learning and deductive learning are two of the most common approaches. Inductive learning is a bottom-up approach， in which learners observe language examples and actively infer grammatical rules. In contrast， deductive learning is a top-down approach， in which learners are first provided with explicit grammatical explanations， followed by examples and practice activities. Nevertheless， both approaches possess their own advantages and disadvantages. Inductive learning can promote deep understandings but it may increase learners’ cognitive load. Conversely， deductive learning can offer explicit instruction and enhance learning efficiency but may reduce opportunities for active thinking. Therefore， this research integrated the strengths of both approaches and developed the Inductive and Deductive English Architecture (IDEA). Specifically， inductive learning was implemented through reading electronic articles， which allowed learners to explore grammatical rules independently. Conversely， deductive learning was supported through a pictorial grammar book， which explicitly explained grammatical concepts.
The IDEA provides learners with multiple grammar learning resources. On the other hand， learners also have diverse characteristics. Thus， individual differences should also be considered in instructional design. Among various individual differences， cognitive styles have been recognized as an important individual difference because they influence how learners process and organize information. One well-established cognitive style classification distinguishes learners as Holists and Serialists， who differ in their approaches to information processing.
Accordingly， this study aimed to investigate the effects of the IDEA on English grammar learning from a cognitive style perspective. More specifically， two empirical studies were conducted at a postgraduate institutes of a university in northern Taiwan. In Study 1， the IDEA 1.0 was developed by integrating electronic reading and the pictorial grammar book. An open-ended questionnaire was applied to explore the learning perceptions of Holists and Serialists. Based on the findings of Study 1， the IDEA 2.0 was further implemented through improvement to the interface design， learning process， and feedback mechanisms of the IDEA 1.0. In Study 2， quantitative analyses and lag sequential analysis were employed to investigate the effects of cognitive styles on English grammar learning in the IDEA 2.0 context.
The results from Study 1 indicated that most learners thought that the IDEA 1.0 could facilitate intuitive understandings of abstract grammatical concepts. Thus， their learning motivation and engagement could be enhanced. However， differences were also observed between each cognitive style group. Holists reported fewer negative perceptions， valued interface aesthetics and interactivity， and preferred intuitive operations and exploratory learning. In contrast， Serialists reported more negative perceptions， were more sensitive to structural organization and details， and required clearer instructional guidance.
The results from Study 2 covered four aspects， which are listed below.
Test Performance: No significant differences were found between the two groups in pre-test scores， post-test scores， or gaining scores. Furthermore， both groups demonstrated significantly higher post-test scores than pre-test scores. This finding suggested that the IDEA 2.0 effectively supported English grammar learning， regardless of Holists or Serialists. Task Performance: No significant differences were observed between Holists and Serialists in task scores， article reading time， grammar reading time， or overall learning time. In addition， Holists performed similarly on both the article-based test and grammar-based test. In contrast， Serialists achieved significantly higher scores on the grammar-based test than on the article-based test. Behavior Frequencies: Holists and Serialists exhibited different learning preferences while interacting with the IDEA 2.0. Specifically， Holists tended to spend more time browsing articles provided by inductive learning whereas Serialists preferred to read the pictorial grammar book offered by deductive learning. Behavior Sequences: Both groups shared certain similarities， such as reviewing grammar content after completing the test. Nevertheless， different behavior sequences were identified between the two cognitive style groups. Holists tended to take the test immediately after reading grammar chapters whereas Serialists were more likely to browse detailed grammar explanations before taking the test.
Holists and Serialists achieved comparable test performance and task performance， they demonstrated different learning perceptions， behavior frequencies， and behavior sequences. Therefore， this research contributes to the integration of inductive learning and deductive learning through electronic reading and the pictorial grammar book， enabling different cognitive style groups to select learning approaches that match their own preferences while maintaining both test performance and task performance. The findings provide valuable implications and solutions for digital learning designers to accommodate diverse cognitive styles. Furthermore， the proposed integrated learning approach may be extended to other subject areas to support the needs and preferences of learners with different cognitive styles.
