1. Introduction
This paper argues for the unsurprising proposition that the language of a culture influences their philosophy of language. The surprise is that our European culture's blindingly obvious traditional theory of language was such a non-starter for ancient Chinese philosophers. Writing about the same time as Plato and Aristotle, they treated a causal-historical philosophy of language as obvious. In Western thought, the causal theory was a modern, revisionist, naturalist theory. That causal theory of language shaped Classical Chinese philosophy as thoroughly as Greek rationalist linguistic theory did Western thought.
Their theory looks modern to us because our version of that theory emerged in the 20th century as a reaction against traditional Western ‘idea theory’ which shaped Western mind-body dualism. There is a natural world of material objects and an intellectual world of immaterial minds and ideas. What makes Chinese theory of language difficult was not only the inherent complexity of Chinese written characters, but that Chinese thinkers did not present it as replacing our familiar use of mind-body dualism to fund a resemblance or picture theory of the relation of word and world.
Pictures represent (stand for, symbolize, refer to etc.) things. Words represent things. How? Western ‘folk’ theory says by representing the mental thing which pictures the natural material thing. Accepting that ‘folk’ theory entails accepting that the theory is blindingly obvious. The Chinese twist on representation seems problematic because it suggests they don't have minds and ideas. If the paradigm problem of philosophy is the relation of mind (ideas) and body (natural objects), then they are not philosophers.
The puzzle deepens when we appreciate that Chinese theory of their writing treats written words as combining pictures and indicative elements to form meaningful units—the things Western rationalists would class as abstract, mental objects, the meanings of words. Why, with a notoriously pictographic/ideographic folk theory of their written word, were Classical Chinese thinkers not compelled to see how obvious the idea-thought represents object-fact theory is to Western common sense?
Language was as central to their philosophical tradition as it was to ours. But while Western philosophy of language became a dominant field in the 20th century, the Chinese viewed their own theory of their language as pivotal during their Classical period (3rd c. bc), keying their account of humans’ place in the natural world. Both traditions of philosophy had strong currents of deflationary theories of language, that dismissed language as superficial in relation to something more real and authentic. The crucial point of divergence was that Western philosophy's reality was the individual soul or mind (psyche) and Chinese reality was a natural world in which humans cooperated via historically evolved social-political structures.
2. Daos and Laws
We start with one key to the puzzle: the Chinese concept of dào (道 path, way). I call this a Dào in the title, but all Classical Chinese philosophy pivoted around that normative/causal concept. In the earliest historical text, Confucius's Analects, that core concept structured the schema for understanding natural guidance. On the other side of the Himalayas (from India to Egypt) the metaphor with a similar gravitational pull on thought was a social practice: kings making laws (and punishing disobedience). These two different conceptual frames structured our respective philosophies of language. On one side, we had ways of speaking and writing, on the other, someone authored or agreed on rules of grammar and spelling.
The metaphysical outcomes were stark. The Western side had concepts, ideas, and meanings to help explain how to use words but also had intellectual world counterparts of grammatical sentences such as thoughts, beliefs, opinions, and a world built of facts, events, and actions. The norms involved the ‘logical/syntactic’ fitting of grammatical sentences to these sentence-like bits of reality. The Ionians and Greeks discovered laws of logic and the concept of a proof (Pythagoras's theorem) to understand how to guide our use of sentences, and thus how to think and believe.
This Western account of words, minds, and the sensible world we call semantics. It grounds the Western mind-body dualism, the theory of minds and ideas. Our theory of knowledge involves getting the semantic mapping right, representing, picturing the world accurately. Beliefs, thoughts, and the sentences that express them will count as known only if they are true, i.e. map accurately onto facts or events. And our normative goal is not only to make true our beliefs and assertions, but to make true the laws of moral conduct or behavior. Philosophers know to invoke a plethora of other conditions, qualifications, and clarifications of this orientation. It fills two millennia of Western thought around this core Western ‘common sense’ view of knowledge of reality.
Looking at Ancient Chinese thought, a familiar reaction is that it is not philosophy. It doesn't struggle with the relation of mind and body which is the essence of metaphysics. It doesn't have a discipline of logic. It didn't have an obvious counterpart of truth until after the spread of Indian Buddhism from the other side of the Himalayas, so it's not recognizable as epistemology. At most it deals recognizably with ethics but lacks the rest. It is merely wise humans moralizing.
What is less noticed in this clash of philosophical civilizations are the two related contrasts we address here: the absence of the Western rationalist ‘common sense’ ‘idea theory of the meaning of words’, and ‘thoughts’ or ‘beliefs’ as the common-sense mental counterparts of sentences. This latter component of Western folk theory anchors the ‘belief-desire’ theory of human free, rational behavior and its role in the Western puzzle about human free will.
The ancient Chinese masters instead operate with a primary focus on social practice and its contrary, natural guidance, the guidance implicit in the Chinese path (dào) metaphor. The implicit goal of ‘knowledge’ was not of picturing a material reality, but of competence, mastery, and know-how in behaving in real contexts. They didn't believe in the importance of correct belief/thought. They neither focused on sentence-sized units of language, of mind (beliefs, thoughts), of natural (causal) process (events), nor of human behavior (actions). The first sentence of the first complete book of learning, the Confucian Analects (Reference Lauca. 5th c. BC), begins ‘to study and regularly practice, that is the life!’ Confucius’ conception of the good life operates within a seemingly mereological conception of nature and language. There are larger, more inclusive, and smaller parts of nature. There are also parts of language, longer and shorter compositional strings of strokes, characters, phrases, slogans, descriptions etc. They are two separate compositional structures.
Chinese thought focused on the individual word or character. The character could be a part of these strings (whether on scrolls or spoken) and it in turn had parts, pictographic, indicative, or rhyming components that made up the individual characters each designed to fit in its square ‘box’ in the vertical bamboo strips of Classical texts. Characters can be compounded in many ways to represent sums or intersections of things (物 wù natural kinds and 器 qì artifacts, implements, tools) as well as interactions in social behaviors (事 shì affairs, business).
3. Idea Theory of Meaning
One way to query whether traditional ‘idea theory’ is an intuitive, common-sense account of human experience or an optional cultural inheritance is to look at the earliest arguments for it. It turns out common sense is not all that common and usually has an evolutionary history behind it. It emerges most recognizably in Aristotle (via Plato). While the origins of idea theory may lie in a long and obscure pre-history of the languages of the ancient civilizations stretching from India to Egypt that eventually adopted the Phoenician invention of a segmentally phonetic alphabet, alphabetic writing motivated the Classical Greek theory of the intellectual, rational realm of being. It has been around long enough that we routinely take it for granted in our way of teaching each other about language. With the typical Greek genius for elaborating a concept by giving it an argument, Aristotle spelled his theory out this way:
Now spoken sounds are symbols of affections in the soul, and written marks symbols of the spoken sounds. And just as written marks are not the same for all men, neither are spoken sounds. But what these are in the first place signs of—affections in the soul—are the same for all; and what these affections are likenesses of—actual things—are also the same. (Aristotle, De Interpretatione, 16:3, Ackrill (trans.), Reference Ackrill1963, p. 42)
Chinese theory of language is as old as that of the Greek rationalists. Parmenides and Heraclitus to Plato and Aristotle overlap with Confucius to Mozi, Mencius, Zhuangzi, and Xunzi. The Chinese masters implicitly shared a theory of language-world relations that more closely resembled Kripke's (Reference Kripke, Harman and Davidson1972) modern alternative to idea theory – the causal, historical-conventional account of language reference. The role of characters in Chinese theory is widely misconstrued in academic orthodoxy by one side in a long-running debate about how to describe Chinese characters: whether to use the term ‘ideograph’ or to adopt some ‘more accurate’ neologism such as ‘logograph.’
Peter Du Ponceau (Reference Du Ponceau1838) functions like a prophet of the ‘logograph’ side of the debate:
Chinese characters represent the words of the language and are intended to awaken the remembrance of them in the mind, they are not therefore independent of sounds, for words are sounds. It makes no difference whether those sounds are simple and elementary, as those which our letters represent, or whether they are compounded from two or three of those elements into a syllable. There are syllabic alphabets, like that of the Sanskrit and other languages, and it has never been contended that they do not represent sounds. And it makes no difference that the Chinese syllables are also words, for that does not make them lose their character of sounds. But, on account of this difference, I would not call the Chinese characters a syllabic, but a logographic system of writing. This being the case, it seems necessarily to follow, that as the Chinese characters are in direct connexion with the Chinese spoken words, they can only be read and understood by those who are familiar with the oral language. (Du Ponceau, Reference Du Ponceau1838, Appendix, p. 110)
Du Ponceau was channeling other Western theorists of language like Ferdinand de Saussure (Reference de Saussure and Harris1983) and Leonard Bloomfield's (Reference Bloomfield1914) ‘scientific’ principle that empirically confirms Aristotle's view that language = speech, and that written language necessarily represents speech and things in the world indirectly through the medium first of speech, then of ideas. They take the word ‘ideograph’ to deny this fundamental principle of linguistics and claim writing directly (without mediation) represents an idea or concept.
Their remedy ignores a less familiar difference between Western and Chinese ‘folk theories’ of language. Chinese theory of their own language did not say it represented Aristotle's ‘images in the mind.’ They thought characters represented natural objects directly because of their history. They didn't have a mental idea or picture theory of the relation of words and world.
That fact seems to opponents of the word ‘ideograph’ to entail that they did not have a theory of meaning since Western folk theory turns meaning into an object in mental space, a mental construct, conception, or abstract object. This distorts traditional Chinese philosophy of language into a silly version of Aristotle's theory – one where pictures (pictograms or ideograms) represent the mental pictures (ideas, meanings, concepts) which in turn represent objects and, theoretically explain how speech itself has meaning, links to the world. This coloring of traditional Chinese language learning theory is thus saddled with the conclusion that speech represents writing rather than the obvious ‘scientific’ law that writing represents speech.
John DeFrancis (Reference DeFrancis1984) turns the alleged empirical law into a tautology by stipulating a distinction between a full and partial writing system. A writing system is full only if it can represent every expression of the spoken language. This allows him to dismiss other obvious ideographs in Western writing such as 1, 2 and &, @, and $. The alternative for advocates of ‘logograph’ is to treat them as representing the English sound ‘o-n-e’ and ‘d-o-l-l-a-r.’
This is, quite frankly, just the insistence that the Western theory of teaching their languages (idea-thought theory) is right (for all languages) and the Chinese theory of teaching theirs is wrong. Further, it is baldly wrong about Chinese languages. The error is continuous with our practice of calling Chinese a language, rather than a language family, a group of related languages. The real value of Bloomfield's ‘spoken language priority’ dictum is to recognize that Chinese is not a single language, but 40+ (depending on degree of mutual intelligibility) distinct spoken languages. It reminds us that China is roughly equivalent to Europe in the number of related but distinct languages. In Europe, we appeal to writing to help us distinguish between different spoken languages and different dialects of the same languages. In China, continuity with this European practice has led to calling Cantonese and Hokkien dialects rather than languages of Chinese.
There is no ‘Chinese’ language in that sense other than the Classical, literary writing – which copies the spoken form of none of the spoken vernaculars. What Chinese ‘ideographs’ wear on their face is that they are a language of a meta-linguistic community, a family of spoken languages, but neither a language ‘spoken in China’ nor a language of the mind. ‘I have no books,’ for example, 我沒有書 uses the negation form of Mandarin, 吾無書 uses the probably unspoken literary negation (and the literary first-person pronoun) and 我冇書 uses the Cantonese negation form. Other triggers might be the third person pronoun, 佢 vs. 他, linking verbs, aspect markers, and a plethora of common expressions as well as direct and indirect object order. Each could be read in the other's language (except, of course, for the unspoken literary form) as a stilted, awkward, or literary way to speak. It would almost always be unnatural speech – considerably more unnatural than how reading written English essays is unnatural in comparison to giving a lecture in English.
Similarly, logogram advocates note that the translation of, e.g., an English word takes two characters in Chinese and conclude that the individual characters are morphemes not words. Europeans use the conventions of writing spaces to individuate and count words and so we get the contrast of ‘long words’ of German because German compound words omit spaces. How can we criticize Chinese for similarly individuating their ‘words’ with the box-like space within which they write each character (especially when the various Chinese languages typically read the character as a single syllable word)?
If Chinese characters ‘stand for’ Chinese words they do so in the same way English written words stand for (represent) English words – to wit, usually not at all. If I want to refer to an English word, I put it in quotes or italics or otherwise signal that I am referring to the word, mentioning it, not using it in talking about us in a natural world environment. Normally, we use a word or phrase in either script to pick out or draw attention to some proper part of the world, not to a sound. I can, of course, refer to the sound, but the token of ‘can’ eleven words back in this sentence does not refer to a sound. It simply is a written word and if my wife says c-a-h-n and I say c-a-n, we say the same word as was written. If the speech-first argument shows Chinese is logographic, it equally shows that English is logographic. The written word ‘can’ is the English word ‘can,’ not its different spoken variations. The spoken variations are equally tokens of the English word ‘can.’
So it is with Chinese 可 (Mandarin kě Cantonese hǒ can, may). The obvious difference is that European written languages are alphabetic with segmental phoneme structure, and Chinese is not. It's something else and the different neologisms proposed to replace ideographic cannot make it ‘just like phonetically impoverished English.’
Further, the Mandarin and Cantonese pronunciations of the character are words in different but related spoken languages, not dialectal variations in sound. The spoken pronunciations of Chinese characters by speakers of Hokkien, Cantonese, and Mandarin are also words but different spoken words of different languages.
This should make us wonder what sound reasoning would make us so vehemently reject calling Chinese characters ideograms. The parallel with ‘&’ is apt. But DeFrancis (Reference DeFrancis1989) is wrong to conclude it represents the English word ‘and’ any more than it does the German ‘und.’ Nor does the logical ‘.’, the mathematical ‘+’, or set theory's ‘∩’ and ‘∪’, or Chinese 同. We write two of the tokens using a segmental phoneme alphabet, and that is it! ‘Ideogram’ remains a perfectly viable candidate for these other examples of written units. We need not infer that they refer to objects in a mental space envisioned in idea-thought theory. They are words that could be read or re-written in other symbolic or linguistic form and pronounced in various spoken languages.
4. Socio-Linguistic Practices
The Chinese tradition's own account of the characters themselves is of a socio-historical process, not of the implicit psychological history of the child's gathering impressions into an abstract pictographic mental language (mentalese) to facilitate her learning of English from her parents. The typical Western ‘folk psychology of language’ presupposed in the argument for calling Chinese characters ‘logograms’ is one the Chinese example requires us to re-examine, not to enshrine it in an a priori principle.
A typical Chinese introduction to their writing system might start with indicatives (指事 zhǐshì pointing to a social interaction) – characters, like above (上 shàng), below (下 xià), and center (中 zhōng) and 一二三. The obvious parallel of the latter with ‘123’ is a reasonable first start to understanding Chinese theory of language. There is a kind of obvious way to use them in sharing information about our situation in the world. Indicatives are typically the first in traditional lists of the several different structural principles of constructing Chinese characters.
The second traditional Chinese category of their characters are pictograms and the account of them recounts the Chinese myth of culture-heroes who invented language and both the social roles and the conventional practices using them. These culture founders started the chain transmitting (teaching) them to today. These Sage Kings are wise but not omniscient – mortal humans. The central tradition credits the Yellow Emperor's historian, Cangjie (now the name of a character-based computer input keyboard) as the inventor of pictographic characters. He was inspired by a hunter who recognized tracks of animals as marking the path to finding them. He scribbled 
to record information about the sun. It ‘evolves’ as others use it into 
then 日, the current graph for a day, or daily, and semantic component of characters related to the sun like 旦 (dawn, morning etc.). Similarly, 
is the causal-historical ancestor of moon (月 yuè month(ly)).
The Sages then used hundreds of indicatives and pictographs in compound metaphorical characters. The Chinese name for this category, meeting of images/plans (會意 huìyì), is the one we are most inclined to translated as ‘ideograph’. The classic example is combining the two natural sources of light above to form 明 (míng bright, clear, understand, insight etc.).
The translation of huìyì as ‘ideograph’ is apt enough for this category, but we needn't infer that they ‘directly invoke ideas.’ It is because huìyì work structurally a bit the way Lakoff and Johnson's metaphorical gestalts explain meaning. The compound graph relates other concepts in an inferentially linked structure of pictograms and indicatives to simulate the gestalt of a familiar bodily encounter with the natural world. The gestalt space is structured around the concepts that play roles in that bodily situation – ‘getting to first base’, ‘strike three’, ‘way out in left field’.
A core example of how this type of character evokes a situation gestalt is 德 (dé power, excellence, virtuosity), the famous companion character to dào. Its earlier forms 
and 
combine the crossroad-like paths formed between separated rice paddies (a pictogram that is completed in the early versions of ‘walk’ or ‘behave’ 
and 
) with pictograms of the eye and heart. The bodily gestalt of finding and following a path becomes the term for good, skilled, correctly learned behavior. This path metaphor, rather than the sentential metaphor of a threateningly powerful male's verbal command, becomes the Chinese conception of a norm of behavior.
This third type of character structure is the most engaging and emphasis on it provokes opponents of ‘ideograph’ to stress instead the fourth structural type where, rather than a visual or bodily metaphor, the structure combines a pictographic category (e.g., man, fish, tree, grass, water, ice etc.) with a phonetic or rhyming component – ‘the kind of fish that sounds like ‘sam’ or rhymes with ‘mammon’ in some, many or most Chinese vernacular languages’.
Opponents of ‘ideograph’ like to focus on these semantic-phonetic compounds instead. They make Chinese a ‘full’ written language by providing a formula for constructing a character for every word of any Chinese spoken language. They object to focus on the other three categories partly on grounds that a large Chinese dictionary will have far more phonetic compound characters than any of the others. This is an unsurprising observation since there are more species than genera, more types of fish and flowers than two.
Were we to focus on the correct characterization of Chinese written content, the traditionalist could count the ratio of occurrences of the other three on a ‘typical’ page of print. Then the proportion of semantic-phonetic compound type would shrink substantially, especially if we include the times a pictogram or indicative is used as a character component! Nor need we infer that since the component is relevant to its contribution to the information in the character that it could not also be semantically or metaphorically relevant as well. The meaning need not be irrelevant.
However, this is to assume that our concern here is with how to scientifically characterize the language. It isn't. We are interested in how their philosophy of language, which their language obviously influences, further shapes their wider philosophy – their metaphysics, ethics, etc. For this purpose, it is important not to distort Chinese theory of their own language into a pastiche formed by illicitly mixing it with Western idea-thought theory.
5. A Compromise Resolution
Armed with their traditional conventional-historical account, the temptation among Chinese theorists of language to invent another immaterial language of abstractions to explain the relation of sounds to natural objects almost completely disappears. ‘Almost’ because the ancient Chinese thinkers were clearly aware of imagination, memory, planning, and dreaming. These played roles in their accounts of humans choosing and following natural world dàos (paths), but not recruited to explain how language links to the natural world. With a real-life instance of Wittgenstein's reductio of idea theory, (Wittgenstein, Reference Wittgenstein1964) an outward sign-object, they would not need an ‘occult’ sign to serve as the meaning of their sign. They would see easily that the norm of use of the written character is not another occult mental character but a social-historical practice – a dào of the word.
The traditional name of the third type of character, the metaphorical mix, uses the Chinese non-semantic conception of these internal goings-on (意 yí idea, intent). We can plan, imagine, remember and dream many kinds of behaviors including speaking, writing, advising etc. The conventions I practiced as a child gave me a start on navigating with the norms of use in my linguistic community. The relevant picture of meaning would be a map of correct behavior showing, e.g., where to take my turn speaking.
To picture normative guidance and natural information is to map possible paths of behavior. Words (characters or spoken words) act more as signposts that help communicate and share path-choice-relevant information. The signpost could indeed be pictures or picture-like, as road signs are in the modern world. However, it would have been obvious to Chinese theorists that the capacity to recognize the sign for horse (馬), a picture of a horse, and a horse are the same capacity. One does not explain the other. We become competent in using mǎ to call attention to the animal, to the picture and to the pictogram in basically the same way.
Classical Chinese thinkers described the information received from the senses as biàn (辯 distinction, discrimination) of, e.g., light from dark, red from blue, sweet from sour, hard from soft etc. Their psychology could otherwise be characterized by those committed to Western idea theory as ‘naïve realism’. They do not assume the eyes convey mental pictures of mental objects consisting of mental substances painted in subjective inner qualia (properties, attributes) which may (or, for sense skeptics, may not) correspond to the color of natural objects. They say their eyes working normally accurately discriminate red from not-red objects in the world and red and not-red written characters. Their eyes discriminate the shape of a horse from the shape of an ox. When one's eyes are abnormal, e.g., one is color blind, then the eyes fail to discriminate X from non-X. This may be the result of distance or disease.
Mozi said: a blind human can say ‘Bright things are white; dark things are black’ and the keen of sight will not correct them. But if we mix black and white things and invite the blind to pick one out, they would not know how to do it. This is why I say the blind do not know black from white, not because they cannot use the words correctly, but that they cannot pick one out from the other. (Mozi (ca. 4th c. BC), 12:8)
Inferences we routinely draw in English also would not strictly follow. Chinese thinkers needn't and probably wouldn't think of the characters either as ‘meanings’ or as having meanings. They have a history linking them via the experience of some ancient coiners and generations of transmissions and alterations to the paths of present use. Those paths are the norms outlining ways of using the words. The ancient Chinese masters neither felt nor had any need for a separate term with the conceptual role of ‘meaning’ in Western idea theory. A conception of the norms (the dàos) of kě (可permissible or possible) usage are functionally adequate.
Those norms are learned and practiced: ‘learn and regularly practice; is that not fulfilling?’ (Confucius, Analects, Reference Lauca. 5th c. BC, 1:1). No one guesses merely from seeing the character 明, how to use it. We learn that from patterns we ‘see’ in adult usage and experiment within our own practice. The metaphor in the story used to help teach students the structure of the indicatives, pictographs, ideographs, and even the semantic-phonetic compounds assist learning, memory, and understanding.
One line of anti-ideograph attack is that the stories we tell in this way about the structure of the characters are sometimes inaccurate as accounts of the etymology and evolution of the character. With our focus on the ideology taught along with the language, the teaching slogans and assists that help learning, and the ways they impact the wider body of social-political naturalist thought, we needn't spend time worrying about that alleged problem. The teaching techniques that work can continue and those interested in actual evolutionary history can still submit their dissertations.
As Ancient Chinese thinkers appreciated, once a social practice is in place it continues to evolve. Conservative Confucians may want to ‘rectify names’ to bring them back into conformity with the ‘original intent’ of the ancient sages (Confucius, Analects, Reference Lauca. 5th c. BC, 13:3), but Mohists can plausibly urge that any rectification should benefit people today given today's conditions. Elaborating the implicit metaphor in the character can facilitate mastery, but knowing the story is different from mastery. That comes with use in coordination with others in a real environment.
6. Types and Tokens
The key point here is that our concept of a word is vague in ways central to the dispute. A standard philosophical treatment distinguishes between word-type and word-token. In European languages, given that distinction, it is natural to treat spoken and written tokens as tokens of the same word-type. However, once we have that apparatus in hand, we can use it to help think about other linguistic categories, both broader and narrower in scope. We naturally treat differences in written tokens like ‘harbor’ and ‘harbour’ as tokens of the same word, but we could treat them as tokens of a spelling type and pronunciations as tokens of a dialect-pronunciation type.
What Chinese orthography reminds us of is that concepts also have a history within what we can call a concept community. The character in the Analects passage translated above as study, 學, is pronounced hôk in Cantonese, xué in Mandarin, and gaku (Han reading) or benkyoo in Japanese. Each occurrence in its context is a token of one type in all these written languages. It is a token of the same written word type, the written morpheme type. Here we can begin to blend in the role of the semantic-phonetic compounds without jumping to DeFrancis's conclusion that Chinese is simply impoverished phonetic writing. Chinese characters token, in addition to word-types, cognate types (for recognizably borrowed words from other languages).
7. Cognates, Synonyms, and (Correct) Translations
We would treat English ‘book’ and German ‘Buch’ as cognates – words in different languages with the same meaning and derived from an earlier history of borrowing the spoken word-type. Clearly many of the tokens of Chinese characters would token cognate types for many of the different language communities of East Asia. This explains how a token of a written word/character would also token a cognate type. It thus acknowledges the relation of Cantonese hôk and Japanese gaku.
We similarly treat different spoken words within a language as synonyms when they have the same meaning. We could say tokens of the two synonyms token the same concept. Xúe and benkyoo are not cognates, but they play the same role, have a similar dào of use with other words in a concept cluster.
Throughout China's long and varied historical evolution of written and spoken forms of language, the writing system provided an interpretation of different spoken languages in the sense of linking both cognates and synonyms. This explains how Chinese written characters could underwrite a kind of translation among all the vernaculars of the Far East.
In this sense, the graphs do not represent ideas directly. However, functionally they do what ideas do in Western idea-thought theory. They underwrite translation among culturally linked languages that historically had shared words and inferentially linked concept structures like the dào, dé, eye, heart, walking/behavior structure discussed above in the example of meeting of images/concepts (會意 huìyì). When two linguistic communities interact enough to learn one another's theories, they learn these gestalt conceptual schemes, these dàos or patterns of inference linking clusters of concepts.
The characters originally provide a link to translating, interpreting, and understanding other Chinese languages. Japanese, Korean, and Vietnamese borrowed Chinese theories, hence concept clusters and inference patterns among them. If they developed a local approximation of the borrowed Chinese pronunciation, the character could token a cross-language cognate type. If they began to read the character using one of their already existing word types with a similar pattern of inference relations, a similar dào of use, they would be like inter-language synonyms. Correct translations of the same spoken theory in different languages could be (and were historically) written with the Chinese characters.
The spoken word is part of a language characterized as whole, not a naturally representing, phonetically structured atom (that is to say, sound does no magic bonding of word & object). Neither does the Chinese character. Chinese students learn its norms of use as she does other linguistic items, as parts of larger units of language. A word-token (in Europe) is typically also a concept-token and sometimes a cognate token. In the former case, we may token other word-types in that context while conforming to the dào of the language. The other terms share a mapped role in the concept space.
Correct translations would also token the concept while conforming to the dào of the translating language. This does not imply that between more widely separated mega-linguistic communities translation is either impossible or substantially more difficult. What it requires is that easier or more simple-minded translation works better between historically related and co-evolving language families. It is also better for examples like ‘sun’ and ‘moon’ and natural kind terms like ‘fish’ and ‘water’. We can expect other terms, like ‘mother’ and ‘father’, to occur in all languages.
This doesn't follow, however, for historically entrenched theories of distinct cultures, no matter how intuitive and obvious they come to seem to speakers in that linguistic community. Monotheistic creator-God theories, however widespread south-west of the Himalayas, is not a natural kind theory nor such that the structure of natural human societies underwrites expecting it in all languages. The missionaries who came to China searching for the Chinese term for the God in their religion with the inference relations to ‘rational’, ‘good’, etc. should be prepared not to find it.
The missionaries could, still, teach their religion by coining a new compound using existing concept-characters like Landlord of the Sky (天主 tiānzhŭ Catholic translation for ‘God’). Protestant missionaries alternately chose to pick a term from existing Chinese theories with a partially similar inference pattern. The Emperor Above (上帝 shàngdì) was the first ancestor of ancient and traditional Chinese Ancestor Worship who tops the afterlife hierarchy of normative authority whose bottom link is following one's father's dào. They could proceed to teach whoever would listen their theory of this unseen object using other terms (created light and dark, humans in male and female etc.) and expect their followers to start to draw inferences to the other existing terms of Chinese in ways that followed the dào of talking about God in their religion.
They would be disappointed to discover that they could not convince (and convert) any of the literate class (starting with Confucius) who had skeptically abandoned the talk of ghosts that survive in an afterlife social hierarchy and with it the normative authority of any imagined emperor above even earlier. The Zhou Dynasty (10th to 3rd c. bc) doctrine of the mandate of heaven claimed justification over the Shang Dynasty (16th to 10th c. bc) by appealing to the authority of sky-nature (天 tiān sky, the heavens, constant dàos) and dismissing the authority of any humanlike affiliation or endorsement. The natural forces leading to rule are not a promise or agreement, but the way nature unfolds. The paradigm of how nature unfolds are the constant dàos of the heavenly bodies.
The very learned 17th c. Jesuit missionaries in China tried to teach their Chinese intellectual counterparts logic so they could deploy one of their several powerful versions of the rationalist, logical proof of God – the ontological argument. It was taken in the West to show on purely rational, logical grounds that God's existence was necessary – logically necessary. The argument assumes immaterial, abstract ideas are the meanings of words. These ideas compose thoughts, some of which are necessarily true. The conclusion of the argument is ‘God exists’ is true. So, their Mission Impossible was to convince Chinese Confucian philosophers that ‘there is a landlord of the sky’ (有天主 ∃x (Sky-Lordx)).
Suppose you, the Confucian, were to accept the first premise:
P1 Sky-Lord is the epitome of normative value (God is perfectly good)
(e.g., on grounds that sky-nature was the ultimate source of normative authority so if there were an owner of sky-nature he would embody good.)
The Jesuit then wants you to accept the inference (translation) of P1 to its synonym P2:
P2 Sky-Lord has all good predicates to the highest conceivable degree.
Now the Jesuit must teach not only logic, but the Aristotelian theory of subject-predicate structure of sentences and thought. Further, it must include the Aristotelian subject-predicate (substance-property) conception of a thing. Chinese grammar does not require a subject so can't require that there be an underlying substance with properties for something to exist. If we don't have a name for it or words to describe it, so much the worse for our language. What naturally exists, exists. It's our job to fit it into our language.
The Jesuits are not trying to prove what Chinese naturalists already accept – that we may not fully understand or be able to fully describe nature (natural dào) in language. They are trying to prove there is an author-creator of nature – including sky-nature. The Chinese philosopher has no trouble understanding the conclusion, but there is no pattern of inference from P2 to that conclusion following the norms of Chinese written or spoken languages.
Of course, Chinese can be converted to Christianity just as some were historically to Buddhism and Islam. Depending on how many of the related clusters of concepts are given character, character compound or character phrase translations they could come to draw inferences in shared ways. Chinese has had extensive co-evolution with Mahayana Buddhism since the 2nd c. and 真 (zhēn) authentic-natural (as opposed to socially constructed) had already acquired the reality vs. illusion-based sense of ‘true’ before more extensive contact with the European West.
8. Aftermath: Natural Objects and Science
The Jesuits might have concluded that Chinese culture was inherently resistant to logic. Still, it quite readily absorbed logic in the 19th and 20th c. when it came as part of the package with natural science. It was resistant to Western self-conceptions of reason and thought when packaged in the supernatural imaginary of idea and thought theory. Buddhism, for all the intellectual excitement it brought, did not take hold partly because it too was too committed to an image vs. reality conception of philosophy.
While it is true that the science that impressed China also came with gunships and armies, what impressed the earliest Chinese advocates of Western learning and logic was evolution, the scientific account of the dào of life, of our place in a natural, physical world. Traditional Chinese Daoism stressed the unity and continuity of nature, the unity of life with natural process. It was naturally more impressed with evolution and science than the more conservative Confucians.
Its concept of reality was one of a whole with parts that were natural (自然 zìrán self-so) or ‘so of themselves’. Things, natural kinds, are parts of the whole that is nature. Different things unfold in different natural causally possible processes – follow their part of the natural dào. Humans and other living organisms are equipped by nature to be able to distinguish the boundaries and shape of their changing possibilities in the natural world. The theory of evolution helped fill out the account of how this came about without attributing an intention, idea, or plan to the universe or its creator.
Each thing has its dào which is part of the dào of the larger thing of which it is a part. This non-dualist picture of reality was implicit in its traditional picture of human language – different Chinese communities with diverse ways of speaking, spoken languages in the larger Chinese linguistic world which used the literary tools that evolved through the centuries to write and communicate with others in that larger community. The common characters and plethora of different pronunciations were parts of that larger linguistic whole.
If we understand that naturalist picture of language correctly instead of distorting it into an unworkable version of semantic idea theory, it need not matter whether we characterize the package as ideographic or logographic. What we should not do is impose our mind-body, idea and thought metaphysics of language in defining those terms. Chinese language theory did not hold that characters represent ideas directly or indirectly. Their theory played a role in Chinese teaching of their language that is broadly like the role ideas and thoughts play in Western teaching (Aristotle's) of their language. It could work because of the relation of writing to the languages of China that make the ways of reading those characters either cognates and/or synonyms.
‘Logograph’ would also not be objectionable if it were not accompanied by a definition that implies the Western teaching theory of spelling and writing – that characters ‘refer to a word-sound in a spoken language’. If we attend to the full role of ‘logos’ in classical Western rationalist theory, so it embraces discourse, law, logic and -ologies in general, then it would preserve its link to natural science and make it a vehicle of sharing information about natural dào – as multiple translators of Daoist texts have suspected (Carus, Reference Carus1898; Hansen, Reference Hansen1992, Reference Hansen2009; Zhang, Reference Longxi1992).
