Yannick Bruneton

doi:10.63563/jspl.2026.007

Articles

Page Path

Article

Between Empiricism and Pragmatism: The “Puzzle Method” for Teaching Hanmun for Beginners

Yannick Bruneton

Journal of Sinographic Philologies and Legacies 2026;2(1):131-160.
Published online: March 31, 2026

DOI: https://doi.org/10.63563/jspl.2026.007

Paris Cité University (UFR LCAO); École Pratiques des Hautes Études (EPHE)

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (https://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

167 Views
7 Download

prev next

Full Article

Download PDF

Abstract
Introduction
Context and Issues Involved in Developing the Method
Presentation of the “Puzzle Method”
The Text as a Puzzle: Suggestiveness and Interest of the Metaphor
Applying the Principles: The Stage of Research
By Way of Conclusion
Notes
Reference List
Appendix

Abstract

Developed in the context of university teaching in Korean studies for beginners in Classical Chinese hanmun, the puzzle method is used to approach unknown texts with a focus on efficiency, as in an exam where time is limited and the use of electronic devices is prohibited. It synthesizes common practices among readers of Classical Chinese, centered around the concept of a “puzzle.” The metaphor works on multiple levels: the text is the puzzle to be reconstructed, its main lines are the ‘big picture’ that the puzzle represents; the pieces are the words, presenting semantic (image) and syntactic (contours) aspects; the pieces are variously loaded with information to suggest the big picture of the puzzle, so that the search primarily targets the most suggestive Sino-Korean characters hancha in order to be effective. The playful aspect of the puzzle, using information provided by characters already known to learners as a starting point to build a search plan that prioritizes the hancha to be searched for, helps develop beginners’ self-confidence. It develops patience, concentration, and reflection in learners, and dispels the belief that the dictionary is a magic tool that immediately provides the correct solution. The method, developed without the use of AI, is environmentally friendly and does not create dependence on specific technological tools, other than knowledge of how to use indexes to search for hancha in paper dictionaries. It was described in detail in a hanmu teaching textbook published in France in 2025.
Keywords: hanmun teaching, puzzle method, 214 radicals ranking, lexical field, grammatical function word (GFW), horizon of expectation

Introduction

“Another new method for classical Chinese!” some might say. In reality, while there are several textbooks on Classical Chinese available on the European market, particularly in English,¹ prior to spring 2025, there was no textbook for studying Classical Chinese in Sino-Korean reading, as used in Korea, hanmun² 漢文, intended for French-speaking beginners seeking to specialize in “Korean studies”³ Han’gukhak 韓國學, or those with other motivations for learning Classical Chinese that are not directly related to Korea. The publication of a textbook in 2025⁴, the result of twenty years of university teaching experience to beginners in Hanmun as part of a training program in “Languages, Civilizations, and Foreign and Regional Civilizations” (LLCER),⁵ therefore clearly fills a gap. The book is neither an exhaustive grammar, nor a collection of grammatical function words,⁶ nor a collection of texts;⁷ it focuses on a method of text analysis in which texts are considered as learning units.⁸ As learner autonomy is the main objective, this can only be achieved through the ability to describe phenomena and analyze a text grammatically. Grammar therefore plays a central role. All learning texts are thus subject to rigorous and systematic grammatical analysis.

In addition, the textbook proposes a method for approaching unknown and uncontextualized texts: a method we have called the “puzzle method,” which is designed to be effective and “ecological.” The latter is intended for use in a time-constrained context, without the need for specific technology, such as during a university exam to assess knowledge. In reality, this method can be generalized to other contexts and audiences, hence the interest in describing it here. The purpose of this article is to present the puzzle method, which has been little theorized due to its empirical and pragmatic nature.

First, we will explain the context in which the method was developed and the issues involved, then we will present the principles and application of the method as concretely as possible.

Context and Issues Involved in Developing the Method

First and foremost, it is useful to briefly review the status of Classical Chinese in contemporary Korea in order to understand how it can be taught to beginners from cultural areas other than Korean.⁹ In short, after having been a “classical” language, i.e., a normative language linked to power in the broad sense and used in multiple registers of literate social life and omnipresent in the fields of knowledge and religion, has now become a “language of culture,”¹⁰ stripped of its binding political dimension, but whose study remains necessary, if not indispensable, in Korean studies, as well as for understanding East Asian civilization.

It is well known that at the turn of the 20th century, the Great Reform of 1894¹¹ led to the decline of hanmun. Its political, literary, cultural, and social decline became effective mainly through the adoption of two measures: 1) the abandonment of hanmun as the official and administrative language written in Chinese characters, replaced by the Korean language written in the Korean alphabet, ŏnmun¹² 諺文, developed in 1443¹³ and called han’gŭl from the 1910s onwards; 2) the abolition of the civil service examination¹⁴ or kwagŏ 科擧 (literally, examination by specialist field), as a consequence of the previous measure. Such a downgrading was politically justified by the official abandonment of Confucianism as the dominant ideology, thus breaking the centuries-old (if not millennia-old) institutional links between the use of hanmun and the functioning of the State.¹⁵

Such measures inevitably had an impact on educational practices, the first stage of which was learning hancha 漢字 (Sino-Korean characters), which was often inseparable from learning hanmun in the form of anthologies, the most emblematic example of which in Korea was the Ch’ŏnjamun¹⁶ 千字文 (literally, “Text of a One Thousand (Different) Characters”), within the framework of the sŏdang¹⁷ 書堂 (local private school). However, before the establishment of the two Korean Republics in 1948, and following the historical experience of Japanese Annexation¹⁸ (1910-1945), the respective linguistic uses of hanmun and Korean, as well as their respective writing systems, remained for half a century in a gray area, called kuk’anmun 國漢文 (between 1895-1948), where the two languages (diglossia)¹⁹ and two writing systems (digraphia) coexisted in a non-standardized manner.

Despite such a language policy, which has been implemented and pursued in both Koreas to this day, the Korean language is clearly characterized by a phenomenon of lexical diglossia,²⁰ with a strong asymmetry between the proportion of the indigenous or hereditary lexicon koyu’ ŏ 固有語, and the Sino-Korean lexicon hanch’a’ ŏ 漢字語, in a respective proportion of about one-third to two-thirds for everyday vocabulary, with a more marked asymmetry for scholarly language. This asymmetry is reinforced by the fact that it is no longer structurally possible to envisage replacing one lexicon with another, a process that some have called “purification of Korean”²¹ kug’ŏ sunhwa 國語純化 or urimal tadŭmgi 우리말 다듬기, because attempts to do so in the 20th century in both Koreas have failed.²² However, the “invisibility” of hancha in public spaces in both Koreas,²³ a consequence of the official adoption of the “exclusive use of han’gŭl”²⁴ 한글전용, implying the “exoticization”²⁵ of hancha 漢字異國化, combined with the common public confusion between language and writing, often suggests the false (and naive) idea that Sino-Korean words have “disappeared” and that it is therefore no longer necessary to learn them, when in fact they are merely hidden under the mask of han’gŭl.²⁶ On the contrary, and paradoxically, it seems that during the second half of the 20th century and up to the present day, within an information society, the predominance of Sino-Korean vocabulary has logically increased in the media due to its semantic efficiency, which also allows for mutual understanding throughout East Asia.

In specialized fields of knowledge and in scholarly language, the predominance of Sino-Korean vocabulary and its ability to adapt to current events by creating new words cho’ŏ’ryŏk 造語力 makes knowledge of hancha essential for mastering the language, which is one of the primary goals of those specializing in Korean studies. However, the principles of Sino-Korean lexicon formation are consistent with hanmun grammar,²⁷ making knowledge of its syntax and the basics of its functioning essential for understanding and memorizing hancha’ŏ. In addition, the use of hanmun, which was widespread in the literate world throughout East Asia until the early 20th century, explains its use as a language of international culture on a regional scale, giving access to the reading of immense corpus of texts, in addition to the fact that Classical Chinese was a major language of translation through which Buddhist culture²⁸ was introduced and widely disseminated, becoming one of its canonical languages.²⁹

More recently, at the beginning of the 21st century, in Europe, following the decline or even cessation of hancha and hanmun teaching in certain universities specializing in Korean studies, and its harmful consequences, there has been a growing awareness of the importance of learning these languages. In this regard, the Premodern Korean Studies in Europe conference held in 2021 provided an opportunity to assess the state of hanmun teaching and to observe very contrasting situations in different European countries, from England to Russia. In some countries, such as Russia, Germany, and France, the teaching of hanmun has been maintained or even strengthened in Korean studies programs, albeit at varying levels and with varying degrees of mandatory status, targeting more or less select audiences.³⁰ In other countries with a more liberal tradition, the situation has deteriorated (the Netherlands, the United Kingdom) or is still developing.

The fact remains that hanmun must continue to be taught as a mandatory core subject in Korean studies, even though only a minority of students specialize in hanmun at the Master’s level, and all teachers are well aware of this phenomenon. Knowledge of the basics of hanmun allows students who want to specialize in ancient, medieval, or pre-modern Korea to avoid making choices by default due to their lack of knowledge of hanmun, the language in which most written sources are written. In general, knowledge of the basics of hanmun is necessary to read sources prior to the founding of the two Republics in 1948.

In Paris, at Paris Cité University,³¹ formerly known as Paris 7 – Denis Diderot, founded in 1970, a Training and Research Unit (UFR) on East Asia was established at the same time as the university, inheriting Korean language courses from the Sorbonne, which began in the second half of the 1950s.³² The first teachers and researchers specializing in history and literature had a solid background in sinology and Classical Chinese,³³ which was a compulsory subject focused on translation. However, for more than half a century, no hanmun teaching textbooks based on their learning had been produced,³⁴ and foreign textbooks were used in addition to the Samguk sagi 三國史記 (Historical Records of the Three Kingdoms, 1145) and Samguk yusa 三國遺事 (Memorabilia of the Three Kingdoms, ca. 1285?). The teaching of Sino-Korean characters began to be imposed there about fifteen years ago as mandatory from the first year of the Bachelor’s degree (L1) in an LLCER-type program (see above). Introducing the memorization and writing of hancha at this level of learning facilitates the memorization of Sino-Korean vocabulary and allows students to accumulate knowledge of several hundreds of hancha, making it possible to learn hanmun in the third year (L3) over two semesters without too much difficulty. In fact, before considering the discovery of hanmun in the best conditions, prerequisites are necessary, if not essential.

Prerequisites for Learning hanmun

There are two main types of prerequisites for learning hanmun. The first, which is linguistic in nature, involves memorizing several hundred basic Sino-Korean characters and the principles of how the language works. The second, which is extralinguistic in nature, involves knowledge of general culture relating to Korea and, by extension, East Asia.

Experience shows that it is difficult for a beginner to start learning hanmun without a minimum stock of more than 200 known characters, ideally more, and provided that this corpus corresponds to specific content: level 7 of the National Sino-Korean Character Proficiency Test,³⁵ Hancha nŭngnyŏk kŏmjŏng sihŏm 漢字能力檢定試驗 corresponding to 150 hancha, to which must be added the 214 radicals from the Kangxi Era Dictionary,³⁶ the Kangxi zidian 康熙字典 (1716). If you don't know all 214 Kangxi radicals, you are required to know half of them,³⁷ provided that they are linked to as many Sino-Korean characters as possible. This is why I suggest memorizing the radicals, not by their classification in ascending order of number of strokes, as is commonly found in dictionaries, but by “frequency,” i.e., by descending order of the number of characters associated with the radicals³⁸ (see Appendix 1 of this article). For example, the three most “frequent” radicals are water (물 水), tree (나무 木), and hand (손 手), ranked 85, 75, and 64 in order of strokes. The 20 most frequent radicals can be used to classify approximately 50% of the hancha in a dictionary containing a total of more than 11,000 characters.³⁹ This not only makes learning more effective, but also more rewarding for the learner because the probability of identifying the radicals learned in a text, either isolated or in combination with other characters, is maximized. Otherwise, the proportion of unknown sinograms in a text is so high that the time and effort invested in looking them up takes up a disproportionate amount of learning time and is a legitimate cause for discouragement. That said, it is normal for learning hanmun, in its broad grammatical sense, to involve discovering and memorizing previously unknown sinograms as new texts are encountered.

In addition to knowing a minimum stock of hancha as defined above, it is necessary to master the various techniques for searching for sinograms in standard dictionaries, which usually include several types of indexes: by the radical 部首索引, then by increasing number of strokes in addition to the radical, by the Sino-Korean pronunciation 字音索引, by the total number of strokes 總劃索引; without relying on digital technology, which is economically and environmentally costly and creates discrimination between learners. Indeed, digital applications developed for automatic Sino-Korean character recognition allow beginners to write unknown characters without taking into account the rules of orderly stroke writing (i.e., writing hancha any which way), which, in practice, has disadvantages in terms of learning to write sinograms themselves and the corollary of being able to read different calligraphic styles. This is because following the rules for writing strokes is the best way to ensure that characters are graphically consistent and legible in different styles, whether cursive or not. From a time management perspective, searching by radical (and the number of strokes in addition to the radical) is the most preferable of the three methods, even if it is not necessarily the fastest, which is searching by pronunciation, because 90% of them fall into the ideophonogrammic category (hyŏngsŏngja 形聲字), which allows the pronunciation to be recognized through the use of the phonetic series⁴⁰ ŭmbu 音符. From this, it is clear that learning hanmun must be based on a solid foundation in hancha.

Another essential linguistic prerequisite is knowledge of the principles of Classical Chinese, including lexical formation.⁴¹ It is not possible to begin reading texts or even sentences without knowing the basic syntax: hanmun is an “SVO” (subject-verb-object) language, where the modifier (or determiner) precedes the modified⁴²(or determined). In the case of isolating languages such as the Sino-Tibetan family to which hanmun belongs, syntax is more strictly adhered to than in other language families. Another fundamental characteristic is the use of grammatical function words (GFWs), which in the history of the Chinese language were first called zhuzi 助字 (“auxiliary characters”), then xuzi 虛字, or xuci 虛辭 (“empty words”). The collection and glosses of GFWs came late in the history of the language,⁴³ and there is no official, definitive list. Moreover, it would be useless to try to establish an exhaustive list for many reasons, the main ones being the diversity of their use, the use of borrowings (mainly phonetic and sometimes graphic), and their combination or use in series. In any case, there are several hundred of these basic function words (between 200 and 500 according to specialized dictionaries), and memorizing them is one of the main challenges in learning the language.⁴⁴

In addition to knowledge of several hundred hancha characters and the principles of hanmun grammar, it is necessary to consider the acquisition of general knowledge about East Asia and Korea, in the case of Korean studies, as another essential prerequisite. Learning about history, geography, literature, and ideas is a decisive asset in situating (in time and space) and understanding the content of texts and their significance, especially when we consider that Classical Chinese was, in East Asia, a classical language, that is to say, a normative language and a tool of expression in the fields of knowledge, politics, academia, literature, and religion throughout the imperial period. Ideally, it is desirable to learn the codification of written culture represented by the “literary genres”⁴⁵ munch’e 文體 in conjunction with knowledge of grammar, even though reference works in Western languages are rare about this topic.⁴⁶ Indeed, East Asian written culture in hanmun in the pre-modern era is fundamentally a literature of genres, knowledge of which provides a basic framework for analysis. As an indication, the great anthology of Korean written culture from the early Chosŏn period, the Tongmunsŏn 東文選 (1478), lists a total of 55 literary genres.⁴⁷ A serious education in Classical Chinese, Korean language, and Korean civilization would need to include knowledge of literary genres along with a deep understanding of grammar.⁴⁸

Presentation of the “Puzzle Method”

The puzzle method⁴⁹ was developed empirically and pragmatically in the university context of teaching hanmun at the third-year undergraduate level (L3) to beginners in this ancient language, who had begun learning Sino-Korean characters in L1 and had a knowledge base of several hundred hancha.

This method aims above all to respond to the three constraints imposed by university exams: 1) the constraint of limited time;⁵⁰ 2) the constraint of reading an unknown text as an assessment exercise;⁵¹ 3) the constraint of not having any electronic devices at one’s disposal, except, possibly, a paper Chinese-Korean dictionary⁵² hanjasajŏn 漢韓辭典 or okp’yŏn 玉篇. In reality, these three types of constraints do not only apply in an academic setting, as they are frequently encountered by readers of hanmun in general, meaning that the method can be generalized and used to advantage by a wider audience.

The method does not claim to use entirely original techniques or special discoveries, but rather relies on common and proven techniques, often intuitive without having been objectified, which we believe are used by all readers of Classical Chinese at different levels. Its main advantage is that it crystallizes and brings together, around a single playful concept⁵³—the “puzzle”—empirical practices designed to be effective and time-saving.

In the textbook, the puzzle method is applied to “false beginners” because it is used during a second major learning phase corresponding to the acquisition of the prerequisites defined above. It follows the word-by-word grammatical analysis of a Korean edition of an introductory text entitled Saja Sohak 四字小學, the “Little Study (in regular groups of) four characters,” according to the version edited by the Chŏnt’ong munhwa yŏn’guhoe (1989), constituting an anthology of extracts from the neo-Confucian corpus of the Four Books and Three Classics 四書三經, belonging to the genre of monggusŏ 蒙求書 (literally “texts sought by the ignorant”).⁵⁴ The text is organized into pairs of groups (or couplets 雙句, 聯) which, in my opinion, constitute its “phrastic unit,”⁵⁵ the majority of which (about two-thirds) are written in parallel style. Preliminary study of this Confucian anthology, which focuses on the practice of filial piety, provides a foundation of knowledge of GFW and a vocabulary of approximately 400 (391) different sinograms, constituting a solid starting point before embarking on the reading of unknown texts.⁵⁶

Regarding the method, we should probably begin by questioning the validity of using this puzzle metaphor. Why talk about a “puzzle”? How can such a metaphor be a source of inspiration for a method of reading and investigation?

The Text as a Puzzle: Suggestiveness and Interest of the Metaphor

A puzzle is defined as “a game of patience consisting of a large number of irregularly cut pieces that must be assembled to reconstruct an image.”⁵⁷ In many ways, the metaphor of a puzzle seems relevant to describe the work of decoding an unknown text in Classical Chinese.⁵⁸ Among the most immediate aspects are the ideas of “game” and “patience.”

The playful aspect of the method lies primarily in its positive approach, whereby unfamiliarity with certain Sino-Korean characters in an unknown text is not considered a handicap, but rather a difficulty to be overcome using known characters. As a result, the method involves developing learners’ self-confidence because it builds on their existing knowledge of hancha and the principles of how the language operates. In the same way that when you start a puzzle, your first instinct is not to lament the disorder, but to find solid and easily identifiable reference points to begin the game, with the goal, if not the resolution, of completing the game and giving yourself the time necessary to do so.

The second key word, “patience,” illustrates the fact that reading a text in Classical Chinese involves a certain relationship to time, which is necessarily different from that of a text written in an alphabetical script, which cannot be one of immediacy, often associated with primacy.⁵⁹ Indeed, it requires time and concentration, it requires reflection: in a word, a certain patience. In addition to patience being a notoriously difficult virtue to acquire (especially in the stressful situation of an exam), it should also be interpreted here as an incentive to break with the belief that a dictionary is a “magical” tool that instantly provides “the” solution, the meaning appropriate to the context of the text being read, and that it is enough to simply consult it. Such a belief, quickly contradicted by the facts, explains why, during written exams where time is limited, students, eager to get started, waste precious time. The rush to find a solution immediately, driven by impatience, often proves to be ill-advised. A dictionary does not instantly provide “the” right solution: it offers solutions, referring to the general problem of the polysemy 多義性 of Chinese characters. Some commonly used sinograms can in fact refer to around twenty different meanings. In these circumstances, before rushing to a dictionary, it is important to remember that it is essential to have at least a vague idea of what you are looking for before embarking on your search, especially if you are not yet proficient in the techniques for searching for hancha in a paper dictionary,⁶⁰ which is often the case for beginners.

What are the “pieces” of the puzzle that make up the text? They are the sinograms and the words they form from one or more sinograms.⁶¹ Schematically, the missing pieces of a puzzle represent unknown characters and words. However, unknown hancha in a text cannot be compared to the “gaps” in a “fill-in-the-blank” text, as these remain visible and contain information that can be partially exploited based on defined prerequisites.⁶²

The pieces of a puzzle contain two types of information: 1) the image 像 they bear, and 2) their particular shape 形, which determines how they fit together with the pieces adjacent to them. Similarly, Chinese characters can be considered in terms of their semantic information or their syntactic aspect, and ignorance of one can, to a certain extent, compensate for ignorance of the other when searching for them in a dictionary. In this vein, and if we stick to the shape of the pieces, it is common to start a puzzle with the pieces whose rare and remarkable shape easily distinguishes them from the others: the corners and edges. The same applies to GFWs whose specific use and function, and position within a group (e.g., beginning or end),⁶³ require them to be followed by other specific syntactic elements. For example, identifying negation words makes it possible to identify the presence of the verbs that follow them, as these verbs play a pivotal role in groups and sentences. In all cases, the search is based on the immediate proximity of what is known, the contiguity of the pieces, without attempting to follow an arbitrarily pre-established order to recompose the whole. Indeed, contiguous pieces are grammatically and semantically correlated; they have a high probability of co-occurrences konggi 共起 and collocations yŏn’ŏ 連語 that are frequent or deduced from the context by logic or the coherence of lexical fields.

The image of the puzzle to be pieced together is the text as a whole, as a set forming a limited, orderly, and intelligible whole, constructed according to rules whose syntax constitutes the mode of arrangement. However, initially, and with a view to completing the game quickly, it is not directly useful to grasp a text in its details,⁶⁴ but first in its broad outlines, considering its most structuring aspects. The “big picture” or subject of a puzzle can be a landscape, a map, a portrait, a piece of art, etc. It is usually one of the elements of the game available to the player to facilitate completion. However, in the case of an unknown text, the big picture is missing, so it is urgent to form an idea of it that is as accurate as possible. To achieve this preliminary objective, anyone with some experience of puzzles knows that the pieces do not all have the same quality, density, or information content 情報量 in terms of their ability to suggest the bigger picture. If we take the example of an urban landscape, a piece representing part of the sky will be of little help in grasping the main features of the puzzle; similarly, in a portrait, pieces representing the background or hair, which have little contrast and are more or less similar and indistinct, will be less likely to reveal their final position, unlike pieces representing characteristic features of a face (eyes, nose, mouth, etc.). However, learners are often more sensitive to quantitative aspects (the number and proportion of sinograms they do not know) than qualitative ones: the quality of the information provided by the hancha they have identified.⁶⁵

Therefore, we can conclude that quantity is ultimately much less important than quality: knowing a few pieces that contain a lot of information about the big picture is more useful for completing the game quickly than knowing a lot of indistinct pieces or elements that are secondary to the main subject. It is therefore understandable that an economical and therefore efficient use of time consists in prioritizing, for a puzzle, the pieces that contain the most information, and, for a text, the Chinese characters that are most suggestive of the general content, the main narrative lines—in the case of a narrative—equivalent to the “keywords” of a text. These fall under what can be called the “dominant lexical fields” (main themes, 主要意味場) of a text, whose coexistence, articulation, and intersection⁶⁶ constitute the overall framework of a text, its narrative, or its discursive development.

The lexical fields of Chinese characters are primarily associated with the lexical fields represented by their radical,⁶⁷ but more broadly, because radicals can be grouped into generic categories,⁶⁸ with some thirty lexical fields developed by the literary tradition.⁶⁹ It is useful to know these fields in order to better understand their use, particularly in figures of style such as parallelism, bearing in mind that the authors of hanmun texts had these lexical fields in mind when they wrote their works. In this vein, Wang Li (1900-1986) developed the idea of parallelism based on the proximity of lexical fields.⁷⁰ In the textbook, I suggest familiarizing yourself with these lexical fields based on a list produced from a comparison of encyclopedic texts or hancha learning (see Appendix 2 in this article). However, the question of lexical fields should not be approached rigidly by seeking to apply the grid they constitute mechanically, as there is no canonical list.

The dominant lexical fields of a text can be identified according to criteria such as the frequency of occurrence of characters or the frequency of identical, semantically similar radicals that can be associated within a broader lexical field, or according to the meaning of the characters (insofar as they are known), particularly verbs, or even the use of rare characters.⁷¹ In a long text, dominant lexical fields are easier to identify than in a short text, where they may be represented by a small number of sinograms. These fields constitute both a kind of “semantic net (or cloud)” and a “horizon of expectation” 期待展望, facilitating the search for unknown hancha. In the imagination and representations, a lexical field corresponds to a sum of denotations, connotations, and “proxemics” that can potentially encompass dozens of words: nouns and verbs (dynamic and stative).⁷² By cross-referencing several (at least two) dominant lexical fields, we can potentially obtain a numerically significant corpus of words capable of encompassing a significant number of unknown hancha still to be discovered. Thus, the degree of concordance between the definitions provided by the dictionary and these lexical fields constitutes an additional criterion of relevance for research, which is thereby facilitated and made more effective.⁷³

The concept of horizon of expectation is commonly used in literary theory to refer to the reception of works by an audience. Here, we give it a more general and less sociological meaning. By “horizon of expectation,” I mean the subjective and personal construction of the sum of assumptions that a reader makes about the understanding of a text in the discovery phase. The richness, depth, and precision of this horizon of expectation will depend on the reader’s level of knowledge and general culture, explaining the need to familiarize oneself with the concept of literary genre due to its high degree of codification.

Ultimately, it is important to remember that unknown Chinese characters in a text do not have equal informational value, just as pieces of a puzzle do not have equal value in terms of the big picture. This highlights the need to develop a strategy, to construct a genuine research plan that prioritizes the characters that are likely to provide the most information about the main themes and central message of the text, thus conducting the research in several stages. At this point, it seems useful to offer some clear guidelines for applying the principles outlined above.

Applying the Principles: The Stage of Research

I now propose several guidelines for applying the puzzle method in a gradual and rational manner. It consists of several stages, the content of which is detailed here.

Exploratory reading of a text without the aid of a dictionary

The first step is to approach an unknown text without any tools, in a calm environment conducive to concentration. During this “discovery” stage, it is important to read the text carefully, in its entirety (in extenso) and several times, in order to extract as much information as possible. Reading, preferably quickly⁷⁴ 速讀 and selectively, allows the reader to become familiar with the structure of the text, but also, with each new reading, to pay different attention to different elements and aspects of the text as they become clearer, while associating them with each other as they are understood.

In order for the instruction to be applied in a way that is both realistic and allows the reader to feel the benefits, it is necessary to start with short texts of just a few dozen Chinese characters,⁷⁵ most of which are supposed to be familiar to reassure learners, and whose content is preferably narrative, so that in addition to known and predictable syntactic rules, an expected narrative structure⁷⁶ 敍事構造, which is part of the “horizon of expectation,” is applied.

The text, equivalent to the frame of the puzzle, must be legible, if possible in large print with a spacious layout, without punctuation,⁷⁷ but where the groups are clearly marked, possibly with line numbering to facilitate the identification of Chinese characters.

The first step is to identify the Chinese characters known in the text, the dominant lexical fields to which the associated hancha belong, starting with the radical-characters pusu 部首 themselves, and to build a knowledge base from which to develop an initial overall representation of what it represents, its big picture, even if it is approximate and incomplete, thus forging a horizon of expectation that raises questions. The aim is therefore to ask the big questions about the content of the text (time, space, situations, actants, actions, discourse, etc.): where? when? who? what? how? why? It is clear that, given their potential informative value in relation to the big picture, the examination of each “piece” or sinogram cannot be neglected.

Ideally, this phase of the work ends when the learner has fully exploited all the information they could glean from the text based on their knowledge (hancha, grammar, general culture), the relationship between the pieces and their connection to the big picture, and when reading and concentration are no longer sufficient. At this stage, even though each character is important, it may be useful to focus particular attention on certain elements.

Specific elements to consider when developing a search strategy

The objective of the discovery stage of an unknown text in hanmun is to develop, in the absence of a dictionary, a plan for researching unknown characters based on the amount of information they contain in order to grasp the big picture of a text.

Information provided by the title and the beginning of the text

With this in mind, particular attention should be paid to the beginning of the text for a number of reasons, starting with the fact that Classical Chinese, which is economical in its expression, frequently uses ellipsis 省略法, and that, as a result, certain crucial pieces of information only appear explicitly at the beginning of the text, sometimes as a single mention, before being omitted later on, where they remain implicit. This phenomenon largely explains why it is difficult to study hanmun from sentences without context, apart from illustrating limited and specific points of grammar. In this vein, one of the most frequent phenomena observed is the ellipsis of the subject. However, for a Korean scholar already familiar with this type of ellipsis in Korean usage, the phenomenon will not seem so unsettling. Another reason for paying attention to the beginning of a text is that it may include a title that usually provides information about the literary genre to which it belongs, thereby providing a wealth of information about its codification: subject matter, language level, use of parallelism⁷⁸ and other stylistic figures, literary models, references to specific periods linked to specific authors, all of which feed into the reader’s expectations. However, it must be acknowledged that beginners in hanmun rarely have such prerequisites, which are acquired through a more in-depth knowledge of Classical Chinese. Furthermore, from the point of view of the narrative structure of a text, the initial situation of a text provides decisive information for its spatio-temporal description, the themes dealt with, and the possible protagonists. Another element that influences the overall understanding of the text is its visual layout.

Personalized redesign of the layout of plain text

Ideally, although this is not feasible in the context of a university exam, it would be useful for readers to be able to redesign the initial layout of the text they are reading in their own way, based on their understanding of the text after several readings during the discovery phase. This is because the layout does not necessarily facilitate reading or visually highlight noteworthy aspects of the text, especially in the case of an older version (not to mention the issue of group demarcation and punctuation), where the main adjustments mainly concern the hierarchy of titles and character breaks as an expression of respect for authority. This work of rearranging the text has many advantages, and not only for beginners: it highlights the structure (whether narrative or discursive, possible speeches, quotations, etc.), the rhythmic basis (frequency of regular groups), the role of connecting elements (connectors, linking words), stylistic devices such as elements written in parallel style (facilitating understanding of the logic of pairing at work); in general, it promotes the gradual distribution of reading effort and memorization of the text. Such work may naturally involve highlighting certain characters depending on the reader’s level of knowledge and reading assumptions. Ultimately, the importance placed on the visual aspect of the text is justified by the role of syntax in isolating languages, but also by the ideophonographic writing system of Chinese characters. Furthermore, these isolating languages make use of functional words.

Sustained attention to functional words

It is clear that, due to their structural nature, sustained attention must be paid to grammatical function words (GFW). These words are the key elements of the main parts of a text; they organize the syntax of each of the groups that make up the text, highlight phrases, facilitate the identification of verbs (the pivotal elements of a sentence), and enable the identification of syntactic functions. In his preface to Hanmun kyosu ch’ŏpkyŏng, Hwang Pagyŏn 黃博淵 (?-?), lamenting in the first quarter of the 20th century the methods of teaching hanmun and the progress of the “new studies”⁷⁹ sinhak 新學 where it was no longer necessary, seeking a method that was easy to understand and likely to succeed, states: “The letters of the civilized world are plethoric, but their essence lies solely in the method (of construction) of one group (syntagm) kubŏp 句法! However, between two and five characters, the groups are complete.”⁸⁰ In fact, it is not often enough said that the genius of hanmun, which simplifies both writing and reading, lies precisely in the conception of a “rhythmic base”⁸¹ in groups of four or five characters, neither short tan’gu 短句 nor long changgu 長句, which generally allow at most a single “pause”⁸² tu 讀 by segmenting the groups into 2+2 or 2+3. Thus, understanding the role of GFWs in groups constitutes the basic grammatical knowledge on which the reading of texts is based, regardless of the length of these texts.

Beginners do not have extensive knowledge of GFWs, however, since learning grammatical words is partially included in lists of basic characters such as kich’o hancha, some may have been learned as part of this learning process or, more simply, through memorization of the Sino-Korean lexicon, which uses these GFWs, albeit in a very uneven and seemingly random manner.⁸³ Among the GFWs that are essential to identify, the following examples can be noted: negation GFWs, most often immediately followed by a predicate; ending GFWs or linking adverb GFWs, which are placed at the beginning and end of groups or which can be used to mark cases of dislocation 分解法, disjunction 離接化, and thematization 主題化; expression GFWs (performative verbs of speech), preceded by the mention of the identity of speakers or characters in the text, which announce the presence of deictics 直示 (words whose referents depend on the situation of the utterance and which mark its beginning and end: first and second person personal pronouns, determiners, and demonstrative pronouns); GFWs equivalent to conjunctions that make logical connections explicit and imply the presence of verbs before and after them; pronoun GFWs and the need they imply to identify their antecedents. Finally, it should not be forgotten that in hanmun, ellipsis also applies to GFWs, so that the frequent absence of connectors 接續語 can easily confuse beginners whose use of the language (as in French, for example) requires the presence of connectors. To remedy this, learners must acquire the reflex to take into account the intrinsic meaning of verbs and the context in order to deduce the grammatical and logical relationships between phrases and clauses.

The purpose of reading a short text several times during the discovery phase is to build up an expectation of the overall meaning of the text, which determines the order of priority for searching for unknown Chinese characters that need to be identified.

Organizing a plan for searching for unknown hancha in a text

The development of a “plan” for researching unknown Chinese characters is personalized and has nothing to do with “academic” learning, referring instead to the notions of empiricism and pragmatism. It is the result of a process that involved, during successive readings and discoveries, putting forward multiple hypotheses about the meaning of the text, which then had to be prioritized before conducting research using a dictionary.

Proceed in steps, with a small group of Chinese characters to search for

If there are a large number of Chinese characters to search for, it is advisable to proceed in successive phases according to the degree of priority identified, with the idea that information about the big picture takes precedence over detail. It is preferable to limit the search phase to a few sinograms so that the answers found can quickly lead to a review of the whole, an assessment, allowing the contours of the big picture to be clarified and the general interpretation of the text to be reviewed, validating or invalidating the initial hypotheses. The more the search for priority sinograms is associated with specific questions relating to semantics or grammatical function, the more the discovery of the relevant information will stand out among the dictionary definitions, even if some characters may resist investigation and remain unsolvable.

Unknown characters resistant to analysis

Among the sinograms that resist analysis are, for example, rare characters pyŏkcha 僻字 or graphically irregular characters ich’eja 異體字, proper names, or, more difficult to grasp, phonetic transcriptions ŭmyŏk 音譯 of non-Chinese words whose constituent characters are used as “phonograms” p’yo’ŭm muncha 表音文子, exclusively for their phonetic value, making knowledge of their meaning useless, apart from their pronunciation. In the case of Korea, these terms may belong to the vernacular language (sometimes called pang’ŏn 方言, or ŏn’ŏ 諺語) or to untranslated Sanskrit pŏm’ŏ 梵語 terms related to Buddhism. Other difficulties may arise from abbreviations⁸⁴ 略文 or other conventions characteristic of certain literary genres, which are not explained in dictionaries. Encountering these types of difficulties at the beginning of the learning process can be seen as an opportunity, even if they are confusing and temporarily constitute a barrier. It is undoubtedly useful for teachers to quickly provoke these difficulties in order to explain them better, with a view to anticipating them. Such difficulties, which are ultimately commonplace, thus allow students to accumulate valuable reading experience (which can be passed on to others), avoiding time-consuming and useless dictionary searches in the future.

Ultimately, the seemingly “lost” time spent on initial reflection, invested in understanding the overall meaning of the text, is justified in that it leads to finding the right solution more quickly in dictionaries, and cannot therefore be considered a waste of time, but rather a useful investment, despite the feeling of incompleteness or impatience that accompanies the search. A better understanding of the big picture allows us to reconsider our overall understanding of the text, which in turn sheds light on the meaning of unknown Chinese characters and allows us to refine our hypotheses about their meaning, and even the order of priority for searching for Chinese characters, thus creating a virtuous circle of analysis.

An example of how the method is applied

It would be tedious to detail here an example of how the puzzle method is applied, since it is based on an individual’s level of knowledge of Chinese characters, which cannot be generalized in abstract terms. Nevertheless, it seems useful to describe the broad outlines of such an attempt, as I did in the hanmun textbook,⁸⁵ this time in detail, based on a preliminary study of an introductory text containing about 400 different characters (see p.12), thus defining a reference corpus of known hancha. The text entitled Pulsaju 不死酒 (“Alcohol of immortality”)⁸⁶ taken from a collection of humorous stories⁸⁷ from Yayue 雅謔 (17th century?) from the Ming dynasty, contains a total of 50 Chinese characters, half of which are supposed to be known to users of the textbook (see Appendix 3 of this article), which is a very favorable starting point for approaching an unknown text.

During the first stage of text discovery, the analysis begins with general observations: proportion of known hancha, proportion of GFWs, presence of radical-characters; identification of lexical fields partly associated with the radicals to the Sinograms, although diverse in such a short text, but whose frequency constitutes another criterion for assessment. The second stage consists of reviewing, line by line and word by word, starting with the title, the information provided by the known hancha, both semantically and syntactically. Detailed analysis of what is known allows us to formulate various hypotheses about the grammatical construction of the groups or their meaning and their relationship to the main lexical fields of the text. At the end of this initial observation, the assessment consists of specifying the research plan for the remaining unknown characters, starting with the most important Chinese characters based on three criteria: word class, lexical field, and frequency of appearance.

The next step is to search for unknown sinograms representing the highest priority level, in this case four sinograms, by choosing the most relevant and consistent meaning with respect to the hypotheses put forward, from among the different meanings provided in the dictionary, by reviewing the various meanings proposed. The new observations obtained either validate or invalidate the hypotheses and significantly advance the overall understanding of the text. It is then necessary to take stock of the knowledge and formulate new hypotheses, going through the text line by line. Ultimately, despite the limited number of characters that make up the text, this approach breaks down the analysis into four successive stages of searching for unknown characters (3 or 4 per stage) in descending order of priority, where the last four Chinese characters searched for turn out to be those of a proper name of persons (difficult to deduce from a dictionary) and the ending GFWs, knowledge of which is secondary to the main message of the text.

By Way of Conclusion

Ultimately, the method we have called the “puzzle method” is characterized by its empiricism and pragmatism and does not claim to be a scientific model in the field of hanmun teaching, but rather a synthesis of common practices that are generally not objectified. Nevertheless, it has been designed with seriousness and with the constant concern of being useful to the broad target readership of all complete beginners in hanmun starting with a modest stock of Chinese characters (around 20% of the corpus of elementary-level Chinese characters in the mandatory school system in South Korea). It should first be considered as the fruit and reflection of more than thirty years of experience reading hanmun, as a learner taught by numerous teachers in France and South Korea, as a researcher in medieval Korean history, and as a university professor of Korean studies teaching undergraduate students. For teachers of Korean studies in French-speaking countries, the urgent need was to fill a glaring gap in methodology by using linguistics as an effective analytical tool, but without being specialist linguists, without resorting to complex linguistic concepts, and without using AI. As it stands, the method is perfectible, open to criticism, and based on a pragmatic approach in that it rejects theorization or systematic grammatical modeling, which is not relevant for beginners. Ultimately, it is intended to be a source of inspiration for students and teachers, who will apply it in the ways that seem most appropriate and effective for each individual.

Notes

1) Cf. Bibliography (Dawson, Fuller, Rouzer, Vogelsang).

2) The term hanmun is strictly reserved here for Classical Chinese and not for hancha 漢字. In North Korea, the term hanmun is used to refer to the teaching of Sino-Korean characters in texts written in Korean. Furthermore, unlike Chinese studies, hanmun does not distinguish between the ancient written language, wenyan 言文, and the ancient or pre-modern spoken language put into writing, baihua 白話, which has a different grammar. Korean studies are understood in Europe as areal study (foreign civilizations) by the humanities and social sciences, a point of view that necessarily differs from Korean studies as “national studies” kuk’ak 國學.

3) Korean studies are understood in Europe as a areal study (foreign civilizations) by the humanities and social sciences, a perspective that necessarily differs from Korean studies as “national studies” kuk’ak 國學.

4) The existing English-language textbooks are excellent, encyclopedic in nature, but their ambition (difficulty) and the audience of future sinologists they target do not correspond to my experience as a teacher of Korean studies.

5) In France, there are two main three-year foreign language teaching programs: a literature-based program, LLCER, and the “Applied Foreign Language” (LEA) program, which is more focused on careers in Translation, Law, Economics and Communication.

6) The first collections or dictionaries of grammatical function words (here abbreviated as GFW), sometimes still referred to as “empty words” xuci 虛辭, xuji 虛字 by some sinologists influenced by Chinese publications, appeared in South Korea in the mid-1990s, as well as in France; see Bibliography (Fan Keh-li). It seems illusory to attempt to compile an exhaustive list of these grammatical words because of their uses: either isolated, combined, merged, in series or in sequence with others, or borrowed graphically and especially phonetically, cf. Yannick Bruneton, Grand manuel de chinois classique Hanmun (Paris: Armand Colin, 2025), 51. The textbook includes a list of 163 summary sheets to help memorize the uses of GFW in the texts studied, as well as their “equivalents” (214).

7) Text collections are very useful, especially if they are organized in a logical progression according to length or difficulty. Experience shows that one of the difficulties in learning is finding texts of the appropriate length, difficulty, and interest for the required level of learners. For our part, and by way of example, I consider works such as Hanmun Kangdok 漢文講讀 by Kim Hyŏlcho of Yŏngnam University (see Bibliography) to be quite exemplary for their progressiveness and diversity. Hanmun Kangdok is organized into eight parts: 1. Proverbs and sayings 俗談.格言; 2. Anecdotes and fables 笑話.寓言; 3. Educational works and Confucian classics 童求書.儒敎古典; 4. Prose texts 散文; 5. Classical hansi poems 漢詩; 6. Novels 小說; 7. Historical biographies 史傳文; 8. Excerpts from the Hundred Schools 諸子百家. Also noteworthy is Yi Ŭngbaek’s Hanjung Hanmun Yŏnwŏn 韓中漢文淵源 (see Bibliography).

8) We have adopted the approach that hanmun should be learned using texts, however short, rather than using sentences taken out of context from the classical tradition, especially if they are excerpts from texts.

9) While it is historically legitimate that in Korea the image of hanmun is associated with imperial culture and the Ancien Régime dominated by Confucian ideology and rote learning of the Classics, it is pointless for students from other cultural areas to be imbued with such prejudices, which are detrimental to their learning. Furthermore, it makes no sense to train them unilaterally in the classical Confucian tradition because they are not destined to take the mandarin examinations.

10) The language of culture is a dominant or prestigious language, but one that is no longer official.

11) The reform is known as Kab-o kyŏngjang 甲午更張 (literally “adjustment” of the kab-o year), and the decree implementing the adoption of the Korean Alphabet as the official language was passed in 1895.

12) The ŏnmun, known as kungmun 國文 (literally “national letters”) during the imperial era was renamed Han’gŭl 한글 under the influence of Chu Sigyŏng (周時經, 1876-1914) in the 1910s, and the use of the word Han’gŭl became widespread from the late 1920s and early 1930s.

13) Chosŏn wangjo sillok, 1443.12.30.

14) The civil service examination was established at the beginning of the Koryŏ Dynasty (918-1392) in 958; Koryŏsa, 9:27b.

15) Generally speaking, this reform was the result of the decline of China’s political and cultural influence and the end of the imperial model of organization.

16) The work, which dates back to the Liang dynasty in the 6th century, is not suitable for beginners. At the time, Tasan Chŏng Yagyong (茶山丁若鏞, 1762-1836) wrote a detailed critique of it in his Ch’ŏnmunp’yŏng 千文評 (Critique of the One Thousand Characters).

17) The sŏdang were officially abolished in 1918 by the Japanese colonial regime.

18) During colonization, schooling was mandatory, as was the teaching of Japanese, involving the adoption of Sino-Japanese vocabulary written in a mixed script known as kukhanmun honyong 國漢文混用.

19) Diglossia 兩層言語 is the coexistence of two languages in the same territory, generally with different social statuses: hanmun was used by the political and intellectual elites, for whom it was a marker of social status.

20) It seems necessary to clarify that diglossia in Korea only concerns vocabulary and not the functioning of the language. Apart from vocabulary, the uses borrowed from hanmun grammar are marginal.

21) The movement to replace Sino-Korean words with words of Korean origin was reportedly in effect in North Korea between 1964 and 1966. It is important not to confuse the removal of Sino-Korean vocabulary with the updating of usage by recognizing terms that have fallen into disuse due to societal changes, as shown in the “Collection of Documents on the Purification of the Korean Language”, Kug’ŏ sunhwa charyojip 國語純化資料集.

22) In North Korea, hancha teaching was reportedly eliminated from the school system between 1948 and 1953. In South Korea, Chinese characters were removed from school textbooks between 1969 and 1972. After a few years, these educational policies had to be abandoned due to the decline in the overall level of Korean language proficiency among schoolchildren, with all the consequences that this entailed in terms of access to the “language of culture.” In South Korea, this unfortunate experiment led to the establishment of the 1800 Sino-Korean basic characters kich’o hancha 基礎漢字 in 1972. In general, it is interesting to note that one of the theorists behind the abandonment of hancha, hancha p’yejiron 漢字廢止論, Oesol Ch’oe Hyŏnbae (외솔 崔鉉培, 1894-1970), although hostile to the writing of hancha'ŏ in Chinese characters, did not advocate abandoning the learning of hancha. See Ch’oe Hyŏnbae, Kŭlcha-ŭi hyŏngmyŏng (Seoul: Chosŏn kyohaktosŏ, 1947), 92.

23) This is not a linguistic phenomenon involving the decline and evolution of the language due to a lack of users or the introduction of a new language (English?), but rather a political and social construct.

24) The adoption of the exclusive use of the Korean alphabet in North Korea, Chosŏn’gŭl 조선글, seems to have taken place very quickly, around 1949, while in the South, it dates back to 1968.

25) Considering hancha as “Chinese” characters (중국글자) belonging to a foreign language, breaking with centuries of practice. However, not all South Korean citizens share this opinion. In France, the current use of the Latin alphabet does not imply that its users feel under the yoke of the Roman Empire, even though there was an ancient Gallic language (written in Greek or Latin script). It is not fair to compare, as some critics of hancha sometimes do, the use of Latin words in French with that of hancha in the modern Korean lexicon.

26) Such a statement is certainly caricatural, but these are the kinds of prejudices that are common among beginners studying Korean at university, and they must first be deconstructed by describing the linguistic facts.

27) This point is not agreed upon by all linguists. Personally, I have not found any cases where hanmun grammar is insufficient to explain the construction of Sino-Korean words. Even if there were exceptions that prove the rule, this would not detract from the usefulness of knowing the basics of hanmun syntax for learning Sino-Korean vocabulary.

28) The translation of the immense corpus of Buddhist texts, generally from Sanskrit, took about a millennium between the Han and Song dynasties. Initially the language used to translate the Tripitaka 漢譯大藏經, Classical Chinese subsequently became established as a canonical language of Buddhism.

29) The other canonical languages of Buddhism are Sanskrit, Pāli, and Tibetan. More recently, English, as an international language, is becoming one.

30) In France, teaching the basics of hanmun is part of the curriculum for all students, whereas in Russia, for example, it is only taught to a minority of students who are also learning medieval Korean.

31) Name in use since 2020, after briefly being called “University of Paris.”

32) The first South Korea native to teach Korean in France in an academic institution was Professor Li Ogg (1928-2001), who was invited by Japanologist Charles Haguenauer (1896-1976); see Elisabeth Chabanol, Souvenirs de Séoul 2 (Paris: Atelier des Cahiers, 2019).

33) These pioneers of Korean studies in an academic environment that had been undergoing renewal since the 1970s, Li Ogg, Daniel Bouchez (1928-2014), and Marc Orange (1937-2023), insisted that social and academic recognition of Korean studies within the field of East Asian studies required recognition of classical studies. This opinion seems to be confirmed and remains relevant today.

34) This can be explained in part by the fact that in France, educational or popular science publications are not considered research work in the evaluation of teacher-researchers. Responsible for teaching hancha and hanmun since 2004 and continuing to the present day, at the L1 and L3 levels, I have accumulated experience in the continuity of learning from hancha memorization to the basics of hanmun. Experience shows that a minority of L3 students, around 10%, wish to continue with hanmun at the Master’s level. However, among these 10%, a significant proportion pursue their higher education up to the PhD level.

35) The introduction of the National Sino-Korean Character Proficiency Test in 1992 by the Korean Language and Literature Society, Han’gug’ŏmunhoe 한국어문회, and validated by the Ministry of Education, allows reference to corpora of Sino-Korean characters defined by linguists according to their level of usage and difficulty. However, the classification of hancha has the limitation of being purely alphabetical. It would have been preferable to classify them according to the thirty or so lexical fields established by literary tradition in order to facilitate the transition to learning hanmun. In my 2025 textbook, I propose a reasoned list of lexical fields (see Appendix 2 of this article), as well as a classification of Sino-Korean characters in texts by lexical field, partly inspired by scholarly tradition;see Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 91-93. Furthermore, the test unfortunately does not include a hancha writing test. This is undoubtedly due to the absence in South Korea, unlike in Japan and China, of a reference system for the rules of writing Sino-Korean characters. The development of this test would therefore have been a good opportunity to establish these rules. As it stands, South Korean websites feature animations of hancha strokes 劃順보기 that do not correspond to Chinese characters, which complicates their teaching.

36) Some of the hancha are common to both lists. It should be noted that the 150 hancha corresponding to level 7 of the National Test include both radicals (44) and functional words (8) from Classical Chinese.

37) Knowledge of all the Kangxi radicals remains essential for quickly acquiring the skill of identifying the radical to hancha, both for searching for characters in dictionaries and for identifying lexical fields. In addition, the radicals can also serve as phonetic series for ideophonograms, which represent 90% of Chinese characters, and thus help you learn their pronunciation (with a small margin of error).

38) With reference to the results given in dictionaries; cf. Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 781-782 (Appendix 1).

39) In this case, the dictionary used is Unp’yŏng Ŏmun yŏn’guso, New Ace Hanhan sajŏn (Seoul: Kŭmsŏng kyokwasŏ, 1989). The principle is very simple: it involved counting the number of Chinese characters associated with each of the 214 radicals and sorting them in descending order.

40) According to specialists, the number of phonetic series varies between 700 and 800 (858); cf. Léon Wieger, Les caractères chinois (Taiwan: Taichung, 2000), 453-623. In Soksŏng hancha haksŭp sajŏn 속성한자학습사전 (Dictionary for learning hancha according to their properties), the number of phonetic series sori kich’ohancha 소리기초한자 is 1056; cf. Kim Tongju and Kim Chuhyŏn, Soksŏng hancha haksŭp chajŏn (Pyongyang: Paeggwasajŏn ch’ulp’ansa, 2005), 16-53. They allow the pronunciation of hyŏngsŏngja to be identified with a success rate of around 80%, because, as in Chinese, there is a (limited) margin of variation. There are no phonetic series for sanghyŏnja 象形字 or chisaja 指事字, whose pronunciation cannot be identified from their written form.

41) Learning the basic principles of hanmun does not take much time: it is covered in a general introduction of a few hours as part of the required LLCER course at Paris Cité University, which includes the history of hanmun in East Asia and Korea, the history of Classical Chinese instruction, and how the grammar works.

42) Thus, for example, in the Sino-Korean lexicon, where words formed of two characters ija’ŏ 二字語 predominate, the inversion of characters changes the meaning of the determination (except in special cases). The determiner-determined order applies not only to the lexicon, but also to the structure of phrases, clauses, and sentences, as Henri Lamasse has clearly shown: “Summary of the preliminary exposition of principles: the fundamental law of position or rule of the determiner: in Chinese, the determiner comes before the determined term (for words as well as clauses)... When there is a conflict between the etymological grammatical function and the positional grammatical function, the latter always prevails (...)”; cf. Henri Lamasse, Sin Kouo wen ou Nouveau manuel de langue chinoise écrite (Hong Kong: Imprimerie de la Société des Missions étrangères, 1922).

43) With the exception of Erya 爾雅, the Yuzhuci 語助辭 by Lu Yiwei 盧以緯 (?-?), in the 14th century, is one of the first dictionaries or repertoires of GFWs (around a hundred); cf. Zhou Dabu and Chŏng Myŏngsu et al. (trans.), Hun’gohak ihae (Seoul: Tonggwasŏ, 1997), 150.

44) The Grand manuel of hanmun contains 163 GFW entries, to which must be added the 214 equivalent GFWs, for a total of 377 GFWs; see Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 40;600-772. The so-called “equivalent” GFWs generally overlap partially with the meaning and function of other GFWs. It is useful to memorize them.

45) The same term munch’e is used to refer to “style,” suggesting a close link between literary genre and style, i.e., the characterization of genres by the presence of more or less obligatory stylistic devices found in model texts. Min Pyŏngsu, interestingly, proposes a list of model texts in China and Korea for each literary genre; Min Pyŏngsu, Han’guk hanmunhak kaeron (Seoul: Taehaksa, 1996), 321-433.

46) To my knowledge, there is no work presenting literary genres in French, and this should be remedied. In meanwhile, there is an old translation of “Treasures from Ancient Writings,” Guwen zhenbao 古文眞寶 (13th century) that provides an overview of the range of literary genres; see Georges Margouliès, Le Kou-wen chinois, collection of texts with introduction and notes (Paris: Libraire orientaliste Paul Geuthner, 1926)

47) See Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 814-815.

48) In the UPCité curriculum, Korean studies is offered as a compulsory subject at the L3 level, with an ECUE (teaching unit) introducing literary genres entitled “SHS and classical materials,” combined with training in classical language. In the space of one semester, it is only possible to cover a limited number of literary genres, supplemented by Min Pyŏngsu’s synthesis.

49) The puzzle method is explained in detail in the textbook and illustrated with an example; see Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), pp. 325-356. In Chinese, the word “puzzle” is translated as p’yŏngdo 평도 拼圖, while in Korean, only the English term p’ŏjŭl 퍼즐 is commonly used.

50) Exams generally take between one and two hours.

51) The advantage of using an unknown text as an assessment exercise is that, among the various types of exercises, it is one of those that requires the most skills that are supposed to have been acquired during the course.

52) In this way, learning hanmun is always associated with learning the Korean language, which is systematically used as the language of translation in the context of “didactic translation” focused on literalness and facilitating assessment.

53) As part of the Classical Language course at UPCité, the puzzle method is used in the second semester as “tutorial work” (Travaux dirigés). The playful and more informal nature of this method is reflected in the fact that students working in pairs (or groups of three) can discuss their understanding of the text being studied and share the often time-consuming task of searching for unknown Chinese characters in a paper dictionary. Working in small teams, where a spirit of cooperation is required, helps to create a more relaxed atmosphere than a formal lecture.

54) See Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), p. 31. Saja Sohak is presented as a reference textbook for learning hanmun in sŏdang, but an article by Yi Tonsŏk traces its origins back to the 1920s and 1930s in Chŏlla Province, i.e., after the theoretical date of the official abolition of sŏdang. cf. Yi Tonsŏk, “Saja Sohag-ŭi hyŏngsŏng-gwa yup’o,” Tongbang Hanmunhak 82 (2020): 331. The most complete version appears to date from 1932 (1,144 characters) and is included in the textbook for comparison with the version of Saja Sohak published in 1989.

55) This is a working hypothesis, but it is partly confirmed by the structure of the text and its use of reference sources, as the various quotations are generally assembled in couplets. The version of the text in the textbook is a partial reproduction of the version published by Chŏnt’ongmunhwa yŏn’guhoe editor in 1989, in 1,000 characters (125 couplets), cf. Yannick Bruneton, Grand manuel de chinois classique hanmun, (Paris: Armand Colin, 2025), 835-837.

56) From a teaching perspective, after the initial discovery phase, teachers find themselves at a crossroads when using resources from this literary genre: they must choose between a more or less traditional approach based on the content of introductory texts focused on quotations from classical texts. In the textbook, we have chosen to break with Confucian classicism and offer short texts classified by theme, beginning with the theme of humor and incorporating several Buddhist texts or texts related to Buddhist culture; See Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 318-327.

57) See definition from the National Center for Textual and Lexical Resources (CNRTL). The invention of the puzzle is attributed to John Spilsbury (1739-1769), a London cartographer and engraver. He created the puzzle in 1766 for educational purposes to facilitate the learning of geography. The first puzzles were made of wood. Puzzle is the name of the fretsaw (a U-shaped saw). By extension, the word “puzzle” has become synonymous with a problem involving multiple components that need to be put together.

58) The issue is dealt with in detail in six points in the textbook; see Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 328-331.

59) From a philosophical point of view, the concept of “primacy” can be considered as the opposite of “secondaryness,” implying a distancing conducive to reflection. This concept is also used in psychology as one of the three axes of characterology according to Gerardus Heymans (1857-1930).

60) Due to the development of Chinese character search software, learning how to search in paper dictionaries is being neglected. The use of new technologies is not a problem in itself, provided that it does not create dependency and does not hinder the acquisition of basic skills. The ability to quickly find hancha in standard dictionaries, whether paper or digital, is a decisive factor in time management. It is preferable to know how to use the different methods, understanding their advantages and disadvantages.

61) It is well known that the vast majority of contemporary Sino-Korean words consist of two characters, accounting for more than two-thirds of the total. One need only open a Korean dictionary to see this for oneself.

62) The usable elements are mainly the radicals to the characters that constitute, even approximately, lexical fields. These are all the more significant in that they refer to concrete lexical fields.

63) It is important to learn the position of a GFW within a group (句上, 句中, 句末, or 句下), but this aspect of learning is often neglected, if not absent. In the summary sheets for learning GFWs in the textbook, the position within a group is indicated systematically. In this case, I was inspired by Cho Chongŏp’s remarkable Hanmun t’ongsŏk 漢文通釋 (General Explanation of hanmun), in which the position 位置 of GFWs is specified (上, 中, 下); see Cho Chongŏp, Hanmun t’ongsŏk (Seoul: Hyŏngsŏl ch’ulp’ansa, 1975), 384-406.

64) To avoid getting bogged down in details, we therefore recommend practicing a series of “speed readings,” see below.

65) It is possible to get an idea of the big picture, if it is not too complex, from knowledge of less than 10% of the total hancha that make up a text.

66) Once the dominant lexical fields have been identified, it is possible to imagine the logical causes of their coexistence through various scenarios. However, this requires a certain knowledge of East Asian culture in general and Korean culture in particular for Korean texts. In any case, combining these fields makes it possible to clarify, guide, and enrich the reader’s “horizon of expectation.”

67) See Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 778-781 (Appendix 1). I propose a classification of radicals into four major lexical fields divided into fifteen subcategories: 1. living things; 2. energy; 3. inert matter; 4. abstract concepts, distributed very unevenly since the category of living things represents half (51%) of the radicals, and that of inert matter more than a third (38%). In the living category: humans, the body, fauna, flora; in the inert category: raw materials, minerals, liquids, processed materials, processed plant materials, raw plant materials.

68) See classification of Kangxi radicals; see Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 779-781 (Appendix 1).

69) See Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 792-794 (Appendix 3). I propose a list of 27 lexical fields (65 subfields) obtained by comparing four Chinese and Korean sources, divided into five main areas: 1. numbers; 2. sky; 3. earth; 4. human; 5. language. It is interesting to note that there is no fixed terminology to designate lexical fields. For example, we find the term mun 門 (lit. “door”) or chok 族 (lit. “family”) in Tasan.

70) See Sin Yongho, Hansihyŏngsingnon (Seoul: Chŏngt’ongmunhwayŏn’guhoe, 2001), 102-112.

71) For beginners, the concept of “rare” characters is relative; it is more relevant to experienced readers.

72) To make the concept of lexical fields easy to understand, simply propose a theme with a keyword and ask learners to write down all the words they spontaneously associate with that term in different registers and word classes: vocabulary consisting of nouns, but not only that: actions, operations, states, properties, and characteristics. On this subject, in his Ch’ŏnmunp’yŏng 千文評 (Critique of the One Hundred Characters), Tasan developed the idea of “word families” 族 or lexical fields by taking into account three elements: form 形, property 情, and operation 事. The textbook gives the example of the lexical field of “sovereign” associated with a hundred words, in French, hanmun, and Sino-Korean; see Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 336-338; also highlighting the fact that despite commonalities, the lexical fields are different and cannot be cultural replicas. This work is even more effective if it can be done directly in the specialized language, Korean and hancha. However, it is rare for beginners in Korean studies to have a large stock of hancha at their disposal.

73) The metaphor of a semantic net, with meshes of varying tightness, makes sense when it comes to retrieving unknown words searched for in the dictionary according to their proximity to identified lexical fields.

74) Here, “speed reading” does not mean reading as quickly as possible, but reading at a steady pace without systematically stopping at unknown elements in the text at the risk of losing your rhythm, in other words, without going back. Reading is fast because it is selective, since the goal is to identify the overall meaning of the text without getting bogged down in details. The aim is to identify unfamiliar characters that are important to understand in order to grasp the overall structure and message of the text.

75) In the textbook, the length of the unknown texts varies between 50 (the first text) and 321 characters, and their layout varies between 5 and 15 lines.

76) The narrative structure usually consists of four main phases: the initial state; the transforming force; the restorative force; the final state. A similar structure is generally found in the construction of classical poems known as kisŭngjŏn’gyŏl 起承轉結.

77) I do not recommend the use of punctuation as a verification method; I prefer didactic translation into one or more foreign languages, including Korean, in order to grasp the relationships between the Korean language, Sino-Korean vocabulary, and hanmun. Furthermore, punctuation is not as necessary in hanmun—whose syntax is already intrinsically rigorous—as it is in a text in a language such as French, where it is essential for proper understanding or to reduce ambiguity of expression.

78) The genres of mun 文 written in parallel style are generally ancient, produced in an official context, for cultural or memorial purposes, whereas the genres for which parallelism is not mandatory are narrative and discursive genres; cf. Yi Chonggŏn and Yi Pokkyu, Han’guk hanmunhak kaeron (Paju: Pojinjae, 1991), 12; see Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 77-78.

79) The new studies, which referred to Western knowledge, were the counterpart to hanhak 漢學 (Chinese classics). This social phenomenon echoes what was said at the beginning of the article about the decline of hanmun from the second half of the 19th century onwards.

80) The term kubŏp is not reserved for poetic genres in this case. See Chŏng Ik, Hanmun kyosu ch’ŏpkyŏng pu chusŏk (Seoul: Pokyŏng munhwasa, 1986), 5; 天下之文博矣而其要不過一句法而已.

81) The rhythmic base is an aspect of prosody, likely linked to the oralization of hanmun, even though the written language was not, by definition, spontaneously oralized in everyday life; cf. Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 70-73.

82) “What is marked by a dot before the end of a group to divide it and facilitate recitation and reading is called tu ‘pause.’” 語未絶而點分之以便誦讀謂之讀. In the official punctuation practices adopted under the Old Regime in Korea by the Palace Library, Cabinet of Books, 秘書省), pauses were marked by a dot in the middle of the character 讀分則點於字之中間, while those at the end of a group were marked on the side; see Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 33.

83) Generally speaking, when hanmun GFWs compete with Korean grammatical constructions, they are not necessarily used (such as the negative imperative, for example).

84) Abbreviations generally follow a certain logic (first or last characters, first and last characters, etc.).

85) See Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 343-356.

86) Such a title does not refer to a specific literary genre. It suggests the supernatural or magical nature of the narrative, while also evoking Taoist thought and culture, in which the theme of immortality is central.

87) In the textbook, after studying the highly Confucian, moralistic, and sententious Saja Sohak, I chose to break with the classicist approach of this introductory text in order to broaden the perspective and highlight that hanmun is not limited to being the written language of the Confucian school and imperial Chinese culture, but that it is a major global language of communication and translation (alongside Sanskrit, Latin, Ancient Greek, and Persian), conveying a “literature of wisdom” 智慧文學, through, among other things, translations of Buddhist texts; see Yannick Bruneton, Grand manuel de chinois classique hanmun (Paris: Armand Colin, 2025), 318-319.

Reference List

Bruneton, Yannick. Grand manuel de chinois classique hanmun. Paris: Armand Colin, 2024.
Chabanol, Elisabeth et al. Souvenirs de Séoul: destins croisés France-Corée de 1886 à 1950. Paris: Atelier des cahiers, 2019.
Cho Chongŏp. Hanmun t’ongsŏk. Seoul: Hyŏngsŏl ch’ulp’ansa, 1975.
Ch’oe Hyŏnbae. Kŭlcha-ŭi hyŏngmyŏng. Seoul: Chosŏn kyohaktosŏ, 1947.
Chŏng Ik. Hanmun kyosu ch’ŏpkyŏng pu chusŏk. Seoul: Pokyŏng munhwasa, 1986.
Chŏng Yagyong. Yŏyudangjŏnsŏ. Munjip, k. 22, chapp’yŏng, Ch’ŏnmunp’yŏng.
Dawson, Raymond. A New Introduction to Classical Chinese. Oxford: Clarendon Press, 1984.
Fan, Keh-li. Le mot vide dans la langue chinoise classique. Paris: Libraire You Feng, 1991.
Fuller, Michael. A. An Introduction to Literary Chinese. Cambridge, Massachusetts, and London: Harvard University Asia Center, 2004.
Ham Hyŏnch’an. Saja Sohak. Seoul: Chŏnt’ong munhwayŏn’guhoe, 1989. ____. Kyosuyong chidosŏ Saja Sohak. Seoul: Chŏnt’ong munhwayŏn’guhoe, 1999.
Kim Hyŏlcho. Hanmun kangdok. Kyŏngsan: Yŏngnam Taehakkyo ch’ulp’anbu, 1996.
Kim Kich’ang and Kim Ch’angjin et al. Kyoyang’in-ŭl wihan Hancha Hanmun. Seoul: Chŏnt’ong munhwa yŏn’guhoe, 2002.
Kim Tongju and Kim Chuhyŏn. Soksŏng hancha haksŭp chajŏn. Pyongyang: Paeggwasajŏn ch’ulp’ansa, 2005.
Lamasse, Henri. Sin Kouo wen ou Nouveau manuel de langue chinoise écrite. Hongkong: Imprimerie de la Société des Missions étrangères, 1922.
Lyssenko, Delphine, and Nicolas. Méthode programmée du chinois ancien. Paris: Editions Lyssenko, 1991.
Margouliès, Georges. Le Kou-wen chinois, recueil de textes avec introduction et notes. Paris: Libraire orientaliste Paul Geuthner, 1926.
Min, Pyŏngsu. Han’guk hanmunhak kaeron. Seoul: T’aehaksa, 1996.
Rouzer, Paul. A New Practical Primer of Literary Chinese. Cambridge, Massachusetts and London: Harvard University Asia Center, 2007.
Sin Yongho. Hansihyŏngsingnon. Seoul: Chŏnt’ongmunhwayŏn’guhoe, 2001.
Vandermeersch, Léon and Jin Siyan. Manuel de chinois classique. Paris: Edition You Feng, 2017.
Vogelsang, Kai. Introduction to Classical Chinese. Oxford: Oxford University Press.
Wang Li. Hanwu shiluxie. Hongkong: Zhonghua shuju, 1973.
Yi Chonggŏn and Yi Pokkyu. Han’guk hanmunhak kaeron. Paju: Pojinjae, 1991.
Yi Tonsŏk. “Saja Sohag-ŭi hyŏngsŏng-gwa yup’o.” Tongbang Hanmunhak 82 (2020).
Yi Ŭngbaek. Hanjung Hanmun yŏnwŏn. Seoul: Chŏnt’ong munhwa yŏn’guhoe, 2000.
Zhou Dabu and Chŏng Myŏngsu et al. (trans.). Hun’gohak ihae. Seoul: Tonggwasŏ, 1997.
www.cnrtl.fr/.
ctext.org/.
db.itkc.or.kr/.
www.hanja.re.kr/.
www.juntong.or.kr/.
www.korean.go.kr/.
www.nl.go.kr/.
sillok.history.go.kr/.

Appendix

Appendix 1.

jspl-2026-007-app1.pdf

Appendix 2.

jspl-2026-007-app2.pdf

Appendix 3.

jspl-2026-007-app3.pdf

Figure & Data

References

Citations

Citations to this article as recorded by

Cite

CITE

export

Copy Download

Format
XML Download

Download Citation

Download a citation file in RIS format that can be imported by all major citation management software, including EndNote, ProCite, RefWorks, and Reference Manager.

Format:

RIS — For EndNote, ProCite, RefWorks, and most other reference management software
BibTeX — For JabRef, BibDesk, and other BibTeX-specific software

Include:

Citation for the content below
Citation and abstract for the content below

Between Empiricism and Pragmatism: The “Puzzle Method” for Teaching Hanmun for Beginners

J Sinogr Philol Leg. 2026;2(1):131-160. Published online March 31, 2026

DOI: https://doi.org/10.63563/jspl.2026.007

Download Citation

Download a citation file in RIS format that can be imported by all major citation management software, including EndNote, ProCite, RefWorks, and Reference Manager.

Format:

RIS — For EndNote, ProCite, RefWorks, and most other reference management software
BibTeX — For JabRef, BibDesk, and other BibTeX-specific software

Include:

Citation for the content below
Citation and abstract for the content below

Between Empiricism and Pragmatism: The “Puzzle Method” for Teaching Hanmun for Beginners

J Sinogr Philol Leg. 2026;2(1):131-160. Published online March 31, 2026

DOI: https://doi.org/10.63563/jspl.2026.007

Between Empiricism and Pragmatism: The “Puzzle Method” for Teaching Hanmun for Beginners

Articles

Page Path

Between Empiricism and Pragmatism: The “Puzzle Method” for Teaching Hanmun for Beginners

Full Article

Abstract

Introduction

Context and Issues Involved in Developing the Method

Prerequisites for Learning hanmun

Presentation of the “Puzzle Method”

The Text as a Puzzle: Suggestiveness and Interest of the Metaphor

Applying the Principles: The Stage of Research

Information provided by the title and the beginning of the text

Personalized redesign of the layout of plain text

Sustained attention to functional words

Proceed in steps, with a small group of Chinese characters to search for

Unknown characters resistant to analysis

By Way of Conclusion

Notes

Reference List

Appendix

Figure & Data

References

Citations

CITE

Download Citation

Format:

Include:

ABOUT

BROWSE ARTICLES

EDITORIAL POLICY

FOR CONTRIBUTORS