編輯情況

許慎，字叔重，汝南召陵（今河南郾城東北）人。師事賈逵，受古文經學，為馬融所推崇，時人譽稱「五經無雙許叔重」。舉孝廉，歷任校長、太尉閣祭酒。所著有《說文解字》和《五經異義》。《說文解字》始作於漢和帝永元十二年(100年)，前後經歷二十餘年，至漢安帝建光元年(121年)，許慎臥病在床，才由其子許沖進上；而《五經異義》已佚，清人陳壽祺《五經異義疏証》輯註較備。

許慎於《說文解字敘》中指出：「古者庖羲氏之王天下也，仰則觀象於天，俯則觀法於地，視鳥獸之文與地之宜，近取諸身，遠取諸物；於是始作易八卦，以垂憲象。及神農氏，結繩為治，而統其事。」由於古人沒有文字，只用結繩等方法來記事，以致「庶業其繁，飾偽萌生」。及至倉頡造字，才使「百工以乂，萬品以察」。然而，隨著文字的演變，而時人又「雖有尉律不課，小學不修」，加上假借字的普遍，使許多字已失去了本義，或者出現誤用的情況。為了「解謬誤，曉學者，達神恉」，許慎遂作《說文解字》。

《說文解字》的書名，許慎這樣解釋：「倉頡之初作書也，蓋依類象形，故謂之文。其後形聲相益，即謂之字。文者，物象之本；字者，言孳乳而寖多也。」

版本

原書現已散佚，但其中大量內容被漢朝以後的其他書籍引用，並有北宋徐鉉於雍熙三年（986年）校訂完成的版本（稱為「大徐本」）流傳至今。宋以後的說文研究著作多以此為藍本，例如清朝的段玉裁注釋本。

19世紀以來，唐朝中期的寫本《說文解字》木部殘卷和口部殘葉被發現，它們是目前可見的最早的《說文解字》版本。

體制

東漢許慎撰著的《說文解字》，是中國第一部按照偏旁部首編排的字典，也是世界上最早的字書之一。它首立部首排列法，以六書理論解釋字形、字義、字音及其互相關係的漢語字典，開創後世字典編排、查檢的先河﹔保存了上古豐富的文字資料﹔闡發了前人的六書說，並首先運用六書理論分析漢字的形體構造，因形說義，因聲求源，是人們認識、掌握上古語音、詞匯和讀通先秦兩漢古籍的重要工具書。此外，《說文解字》釋字時往往先列出小篆，如果古文和籀文不同，則在後面列出。然後解釋這個字的本義，再解釋字形與字義或字音之間的關係。總括而言，《說文解字》是科學文字學和文獻語言學的奠基之作，在中國語言學史上有極其重要的地位。

內容編排

關於《說文解字》的編排，《說文解字序》說：「此十四篇，五百四十部，九千三百五十三文，重一千一百六十三，解說凡十三萬三千四百四十一字。」由此可見，本書分為十四篇，另加序一篇，共十五篇。篇下開列部首，如第一篇有部首十四個。部首統攝字，有九千三百五十三個﹔字之下加解說，包括意思，讀音等，解說時又附以古文、籀文、俗字、奇字等異體字，稱重文，合計一千一百六十三個。

在北宋時期，徐鉉等人作了校訂。其中在校訂中對原書的編排作了一些改動和補充。包括：

• 把原書十四篇各分上下

• 增加反切

• 增加注釋

• 增加了一些見於古代典籍的文字——新附字。

那麼如此眾多的被解釋字是怎樣有系統地組織在一起的呢？作者的編排原則是甚麼？這裏牽涉兩個編排的原則。第一是關於部首之間的排列次序﹔第二是每部首之下統攝之字的排列方法。這在《說文解字序》裏也有說明：

又說：

段玉裁說：「凡部首之先後，以形之相近為次﹔凡每部中字之先後，以義之相引為次……」

《說文》全書十五篇，正文十四篇，共收正字九千三百五十三個，重文一千一百六十三個，許慎按照不同的表意或表聲偏旁（多數為表意偏旁），把這些篆體漢字分為五百四十部。例如從木（以木為偏旁）的字 ，如楷、梫、桂、棠、杜、槢、樿等，都放在木部。從水的字（以水為偏旁），「河、泑、涷、涪、潼、江、沱」等，都放在水部。如此等等，這就是「分別部居」。

評價及貢獻

歷代不少學者都曾研究《說文解字》，當中尤以清朝時研究最為興盛。段玉裁的《說文解字注》、朱駿聲的《說文通訓定聲》、桂馥的《說文解字義證》、王筠的《說文句讀》尤備推崇，四人也獲尊稱為「說文四大家」。

說文對於文字學貢獻，在於網羅當時許多的小篆、籀文、古文等古代文字，著眼於本形本義，保存當時的文字說解，不僅為第一部文字學鉅著，事實上兩千多年來的文字學便是說文之學。

Circumstances of compilation

Xu Shen, a Han Dynasty scholar of the Five Classics, compiled the Shuowen Jiezi. He finished editing it in 100 CE, but due to an unfavorable imperial attitude towards scholarship, he waited until 121 CE before having his son Xǔ Chōng present it to Emperor An of Han along with a memorial.

In analyzing the structure of characters and defining the words represented by them, Xu Shen strove to disambiguate the meaning of the pre-Han Classics, so as to render their usage by government unquestioned and bring about order, and in the process also deeply imbued his organization and analyses with his philosophy on characters and the universe. According to Boltz (1993:430), Xu's compilation of the Shuowen "cannot be held to have arisen from a purely linguistic or lexicographical drive." His motives were more pragmatic and political. During the Han era, the prevalent theory of language was Confucianist Rectification of Names, the belief that using the correct names for things was essential for proper government. Xu's postface (xù 敘) to the Shuowen Jiezi (tr. O'Neill 2013: 436) explains: "Now, as for writing systems and their offspring characters, these are the root of the classics, the origin of kingly government, what former men used to hand down to posterity, and what later men use to remember antiquity." Compare how the postface describes the legendary invention of writing for governmental rather than for communicative purposes:

The Scribe of the Yellow Emperor, Cangjie, observing the traces of the footprints and tracks of birds and wild animals, understood that their linear structures could be distinguished from one another by the differences between them. When he first created writing by carving in wood, the hundred officials became regulated, and the myriad things became discriminated. (tr. O'Neill 2013: 430)

Pre-Shuowen Chinese dictionaries like the Erya and the Fangyan were limited lists of synonyms loosely organized by semantic categories, which made it difficult to look up characters. Xu Shen analytically organized characters in the comprehensive Shuowen Jiezi through their shared graphic components, which Boltz (1993:431) calls "a major conceptual innovation in the understanding of the Chinese writing system."

Structure

Xu wrote the Shuowen Jiezi to analyze seal script (specifically xiǎozhuàn 小篆 "small seal") characters that evolved slowly and organically throughout the mid-to-late Zhou dynasty in the state of Qin, and which were then standardized during the Qin dynasty and promulgated empire-wide. Thus, Needham et al. (1986: 217) describe the Shuowen jiezi as "a paleographic handbook as well as a dictionary".

The dictionary includes a preface and 15 chapters. The first 14 chapters are character entries; the 15th and final chapter is divided into two parts: a postface and an index of section headers. Xǔ Shèn states in his postface that the dictionary has 9,353 character entries, plus 1,163 graphic variants, with a total length of 133,441 characters. The transmitted texts vary slightly in content, owing to omissions and emendations by commentators (especially Xú Xuàn, see below), and modern editions have 9,831 characters and 1,279 variants.

Sections

Xu Shen categorized Chinese characters into 540 sections, under "section headers" (bùshǒu, now the standard linguistic and lexicographical term for character radicals): these may be entire characters or simplifications thereof, which also serve as components shared by all the characters in that section. The number of section headers, 540, numerologically equals 6 × 9 × 10, the product of the symbolic numbers of Yin and Yang and the number of the Heavenly Stems. The first section header was 一 (yī "one; first") and the last was 亥 (hài, the last character of the Earthly Branches).

Xu's choice of sections appears in large part to have been driven by the desire to create an unbroken, systematic sequence among the headers themselves, such that each had a natural, intuitive relationship (e.g., structural, semantic or phonetic) with the ones before and after, as well as by the desire to reflect cosmology. In the process, he included many section headers that are not considered ones today, such as 炎 (yán "flame") and 熊 (xióng "bear"), which modern dictionaries list under the 火 or 灬 (huǒ "fire") heading. He also included as section headers all the sexagenary cycle characters, that is, the ten Heavenly Stems and twelve Earthly Branches. As a result, unlike modern dictionaries which attempt to maximize the number of characters under each section header, 34 Shuowen headers have no characters under them, while 159 have only one each. From a modern lexicographical perspective, Xu's system of 540 headings can seem "enigmatic" and "illogical". For instance, he included the singular section header 409 惢 (ruǐ "doubt"), with only one rare character (ruǐ 繠 "stamen"), instead of listing it under the common header 408 心 (xīn "heart; mind").

Character entries

The typical Shuowen format for a character entry consists of a seal graph, a short definition (usually a single synonym, occasionally in a punning way as in the Shiming), a pronunciation given by citing a homophone, and analysis of compound graphs into semantic and/or phonetic components. Individual entries can additionally include graphic variants, secondary definitions, information on regional usages, citations from pre-Han texts, and further phonetic information, typically in dúruò (讀若 "read like") notation.

In addition to the seal graph, Xu included two kinds of variant graphs when they differed from the standard seal, called ancient script (gǔwén 古文) and Zhòu script (Zhòuwén 籀文, not to be confused with the Zhou dynasty).

The Zhòu characters were taken from the no-longer extant Shizhoupian, an early copybook traditionally attributed to a Shĭ Zhòu, or Historian Zhou, in the court of King Xuan of Zhou (r. 827–782 BCE). Wang Guowei and Tang Lan argued that the structure and style of these characters suggested a later date, but some modern scholars such as Qiu Xigui argue for the original dating.

The guwen characters were based on the characters used in pre-Qin copies of the classics recovered from the walls of houses where they had been hidden to escape the burning of books ordered by Qin Shihuang. Xu believed that these were the most ancient characters available, since Confucius would have used the oldest characters to best convey the meaning of the texts. However, Wang Guowei and other scholars have shown that they were regional variant forms in the eastern areas during the Warring States period, from only slightly earlier than the Qin seal script.

Even as copyists transcribed the main text of the book in clerical script in the late Han, and then in modern standard script in the centuries to follow, the small seal characters continued to be copied in their own (seal) script to preserve their structure, as were the guwen and Zhouwen characters.

Character analysis

The title of the work draws a basic distinction between two types of characters, wén 文 and zì 字, the former being those composed of a single graphic element (such as shān 山 "mountain"), and the latter being those containing more than one such element (such as hǎo 好 "good" with 女 "woman" and 子 "child") which can be deconstructed into and analyzed in terms of their component elements. Note that the character 文 itself exemplifies the category wén 文, while 字 (which is composed of 宀 and 子) exemplifies zì 字. Thus, Shuōwén Jiězì means "commenting on" (shuō "speak; talk; comment; explain") the wén, which cannot be deconstructed, and "analyzing" (jiě "untie; separate; divide; analyze; explain; deconstruct") the zì.

Although the "six principles" of Chinese character classification (liùshū 六書 "six graphs") had been mentioned by earlier authors, Xu Shen's postface was the first work to provide definitions and examples. He uses the first two terms, simple indicatives (zhǐshì 指事) and pictograms (xiàngxíng 象形) to explicitly label character entries in the dictionary, e.g., in the typical pattern of "(character) (definition) ...simple indicative" (A B 也...指事 (也)). Logographs belonging to the third principle, phono-semantic compound characters (xíngshēng 形聲), are implicitly identified through the entry pattern A… from B, phonetically resembles C (A...從 B, C 聲), meaning that element B plays a semantic role in A, while C gives the sound. The fourth type, compound indicatives (huìyì 會意), are sometimes identified by the pattern A...from X from Y (A...從 X 從 Y), meaning that the compound A is given meaning through the graphic combination and interaction of both constituent elements. The last two of the six principles, borrowed characters (aka phonetic loan, jiǎjiè 假借) and derived characters (zhuǎnzhù 轉注), are not identifiable in the character definitions.

According to Imre Galambos, the function of the Shuowen was educational. Since Han studies of writing are attested to have begun by pupils of 8 years old, Xu Shen's categorization of characters was proposed to be understood as a mnemonic methodology for juvenile students.

Textual history and scholarship

Although the original Han dynasty Shuōwén Jiězì text has been lost, it was transmitted through handwritten copies for centuries. The oldest extant trace of it is a six-page manuscript fragment from the Tang dynasty, amounting to about 2% of the entire text. The fragment, now in Japan, concerns the mù (木) section header. The earliest post-Han scholar known to have researched and emended this dictionary, albeit badly, was Lǐ Yángbīng (李陽冰, fl. 765–780), who "is usually regarded as something of a bête noire of Shuowen studies," writes Boltz, "owing to his idiosyncratic and somewhat capricious editing of the text."

Shuowen scholarship improved greatly during the Southern Tang-Song dynasties and later during the Qing dynasty. The most important Northern Song scholars were the Xú brothers, Xú Xuàn (徐鉉, 916–991) and Xú Kǎi (徐鍇, 920–974). In 986, Emperor Taizong of Song ordered Xú Xuàn and other editors to publish an authoritative edition of the dictionary.

This was published as the 説文解字繫傳 Shuowen Jiezi xichuan.

Xu Xuan's textual criticism has been especially vital for all subsequent scholarship, since his restoration of the damage done by Lǐ Yángbīng resulted in the closest version we have to the original, and the basis for all later editions. Xu Kai, in turn, focused on exegetical study, analyzing the meaning of Xu Shen's text, appending supplemental characters, and adding fǎnqiè pronunciation glosses for each entry. Among Qing Shuowen scholars, some like Zhū Jùnshēng (朱駿聲, 1788–1858), followed the textual criticism model of Xu Xuan, while others like Guì Fù (桂馥, 1736–1805) and Wáng Yún (王筠, 1784–1834) followed the analytical exegesis model of Xu Kai. One Qing scholar, Duan Yucai, stands above all the others due to the quality of his research in both areas. His annotated Shuowen edition (Shuowen Jiezi Zhu) is the one most commonly used by students today.

Although the Shuowen Jiezi has had incalculable value to scholars and was traditionally relied upon as the most important early source on the structure of Chinese characters, many of its analyses and definitions have been eclipsed as vague or inaccurate since the discovery of oracle bone inscriptions in the late 19th century. It therefore can no longer be relied upon as the single, authoritative source for definitions and graphic derivations. Xu Shen lacked access to oracle bone inscriptions from the Shang dynasty and bronzeware inscriptions from the Shang and Western Zhou dynasty, to which scholars now have access; they are often critical for understanding the structures and origins of logographs. For instance, he put lǜ (慮 "be concerned; consider") under the section heading 思 (sī "think") and noted it had a phonetic of hǔ (虍 "tiger"). However, the early bronze graphs for lǜ (慮) have the xīn (心 "heart") semantic component and a lǚ (呂 "a musical pitch") phonetic, also seen in early forms of lǔ (盧 "vessel; hut") and lǔ (虜 "captive").

Scholarship in the 20th century offered new understandings and accessibility. Ding Fubao collected all available Shuowen materials, clipped and arranged them in the original dictionary order, and photolithographically printed a colossal edition. Notable advances in Shuowen research have been made by Chinese and Western scholars like Mǎ Zōnghuò (馬宗霍), Mǎ Xùlún (馬敘倫), William G. Boltz, Weldon South Coblin, Thomas B.I. Creamer, Paul Serruys, Roy A. Miller, and K.L. Thern.