Truth regularly plays an anaphoric role: like pronouns, ‘true’ can be employed as a way of avoiding repetition of language already used. Chatting with her kitten Dinah on a snowy winter day, Alice imagines the trees outside coming to life in the summer, dressing themselves in green and dancing in the wind. She then says:

Here, ‘it was true’ repeats the content of her previously stated imaginings: she wishes that the trees would come to life in the summer, dressing themselves in green and dancing in the wind. Had she not used ‘true’ anaphorically, she would have had to resort to uttering all the same sentences once again.5

It’s commonly observed that ‘true’ is a device of semantic endorsement. That is to say, one can endorse a thought or agree with what it says by means of ‘true’. Semantic endorsement is a specific (and incredibly common) application of truth’s previously considered anaphoric role. Consider this exchange between Alice and the Duchess:

The Duchess here endorses what Alice said. She could have simply repeated after Alice: ‘He might bite’. She could have said ‘Yes indeed’. She could have nodded, or given a thumbs up. All of these actions would have been means for agreeing with Alice’s claim. Because of the anaphoric role of ‘true’, it can be used as a means for endorsement, as predicating it amounts to reasserting what Alice previously asserted.

In another case, the narrator responds to Alice’s thoughts as she falls down the rabbit hole:

The narrator here is communicating that they believe that it’s very likely that Alice wouldn’t grumble about falling off the top of her house now, were she to do so. The narrator is assessing what Alice said, and the combination of ‘which’ and ‘true’ allows the narrator to refer to and then assess the content of Alice’s thoughts without explicitly expressing them all over again.

Another frequently noted feature of ‘true’ is its generalizing ability. For example: all tautologies are true. In asserting that all tautologies are true, we are expressing a commitment to the endorsement of each individual tautology, and without having to individually express each one of them. Consider this admonition from the Red Queen to Alice:

The Red Queen is effectively advising Alice that for any p, she should assert that p only if p.7 She is deploying ‘true’ in a way that lets her generalize over all propositional contents.

If snow is white, then ‘Snow is white’ is true. Hence we can use ‘true’ to say things about the world by saying things about language. By the same token, when ‘true’ is applied to a piece of language, we can “descend” and infer something about the world. If I come to believe that the sentence on the board is true, and the sentence is ‘Tigers tickle terribly’, then I may infer that tigers tickle terribly. Semantic ascent is closely related to truth’s role in producing anaphora and generalizations; in effect, because of the equivalence between <p> and <<p> is true>, such communicative features are possible. For example, the Duchess is effectively semantically ascending when she replies “Very true”; she is predicating truth of the previously spoken content.

These four dimensions of truth form our minimal characterization. Truth is a notion we employ in generating anaphora, engaging in semantic ascent and descent, and generalizing over propositional contents, all of which then allow it to serve as a means for semantic endorsement (and rejection). None of this is news to anyone invested in the philosophy of truth, of course. That is the point of offering a minimal characterization: by articulating what is held in common by all partisan theorists, we latch onto the shared phenomenon that concerns them all. We can then use that characterization to see if that phenomenon was of concern to ancient Chinese philosophy.

Furthermore, anyone who deploys terms that accomplish these four tasks has the concept TRUTH, since that is the concept that gives those terms their meaning. Whatever else they are, concepts play a role in cognition that enables understanding and communication. Which ideas you can understand and express to others is, in part, a function of which concepts you possess. So competence with alethic language is indicative of the possession of the concepts underlying that language. Note that we are offering only a sufficient condition for possessing TRUTH: cognitive competence with the sorts of expressive abilities for which alethic language is distinctively used. In line with the neutrality driving our minimal characterization, this sufficient condition is accepted by those who think possession of alethic language is necessary for TRUTH possession (e.g., Horwich 1998), and those who think it isn’t (e.g., Asay 2013 and Brons 2016).

None of these claims wades into controversial territory as to what it is for something to be true, or what constitutes the nature of the property truth, or the analysis of the concept TRUTH. Instead, the minimal characterization speaks to what people implicitly understand about truth when they encounter and use alethic language in ordinary settings.8 Bear in mind that the minimal characterization is not a theory or definition of what it is to be a truth predicate.9 A truth predicate, for instance, may just be any predicate whose extension includes all and only truths.10 But that definition won’t be helpful for identifying truth predicates out in the wild of natural language. The minimal characterization focuses on the fact that when we use alethic language, we are doing things like generalizing, semantically ascending, and semantically endorsing. These doings are the key to detecting alethic phenomena in other languages—if speakers and writers are doing those same things, then the language they are using to do so constitute that language’s alethic resources.

By focusing on the role of alethic language, our minimal characterization is less theory-laden than has been the norm with our interlocutors. They, too, begin by laying out what they’re looking for when they look for alethic phenomena in the ancient Chinese tradition (see Saunders 2022: 3–8). But their touchstones for truth are far less theoretically modest than ours, and often are committed to various interpretive perspectives about ancient Chinese philosophy. As will soon be clear, our methodology detects alethic phenomena in many of the same places that other theorists have found them. This is unsurprising, since those theorists obviously have a decent handle (however implicit) on how to detect the concept of truth within the languages they understand. A major virtue of our minimalist approach is that it offers a strong rationale for why it is correct to find alethic phenomena there. Other theorists have gestured toward various principles underlying their alethic identifications, but these have been, at best, incomplete and insufficiently neutral. Fraser (2012: 360) highlights the anaphoric dimension of truth, but at the cost of committing to Brandom’s (1994) version of prosententialism. Leong (2015: 69) lists some common features of truth, but mixes together semantic, epistemological, and conceptual matters. Harbsmeier (1989) nods towards Tarski’s (1956) work on truth and its connection to semantic ascent, but it’s left unclear how to then apply Tarski’s framework to the task at hand. Furthermore, Harbsmeier relies on a confusing distinction between “semantic” and “abstract” concepts of truth. To make his case that the notion of truth is alive and well in ancient Chinese philosophy, Harbsmeier provides a number of translations of texts that deploy ‘true’ when translating the Chinese. As he notes, these translations may “all be taken to beg the question” (1989: 141). Indeed, that a term is translated using ‘true’ only pushes the question back a step: is it appropriate in that context to translate using ‘true’, and why? The minimal characterization of truth provides the correct answer: if the authors of the texts are doing the same things we are doing when we use ‘true’, then it is appropriate to use ‘true’ to translate the words that perform those same actions.

In general, ran 然 expresses some event or state of affairs as “so” or “being the case.” For example, the early Chinese word for “spontaneity”, an important term in Daoist philosophy, is ziran 自然, meaning, “so of itself” or “being the case on its own”. Ran also occasionally functions as an affirmative response to a question, as in the Analects when someone named Jie Ni asks Zilu if he is a student of Kongzi (Confucius), to which the latter responds “ran” or “yes”. However, ran is also deployed in semantic contexts to affirm or reject an utterance, in turn functioning as an MTP, and we discuss a handful of representative cases in this section. Before continuing, however, we should say that scholars have already noticed ran’s affinities with truth in semantic contexts. Leong Wai Chun, for example, offers a detailed discussion of ran and truth, highlighting four features that it exhibits in alethic contexts: its application to linguistic units, its accordance with Tarski’s “Convention T”, its role in making generalizations, and how it enables speakers to justify, endorse, or reject doctrines (2015: 69). Other discussions of truth and Chinese philosophy also treat ran as perhaps the closest approximation of truth the Chinese tradition has to offer (Van Norden 2007: 372–373; Fraser 2020: 121–123). And even prior to philosophical scrutiny, some sinologists readily helped themselves to translating ran as “true” in a diverse array of early Chinese literary contexts (Harbsmeier 1998: 193–209; Roetz 1993).

While we agree with these scholars regarding ran’s proximity to truth, we also note that they form their conclusions without employing either the tripartite distinction or the minimalist characterization. For example, Leong claims that “[i]n English, the concept of truth has” four features (Leong 2015: 69), but the features he lists conflate linguistic behavior with logical and epistemological features of truth. (Nor is the concept of truth something that belongs to English.) This leads him to ascribe a substantive concept of truth to early Chinese thinkers, whereas we believe that these features only serve to identify truth predicates within ancient Chinese texts.11 Similarly, Van Norden’s approach to ran relies on a “logical” conception of truth according to which one has a thin conception of truth “just in case they have a concept that adequately corresponds to ‘true’ in the following schema: ‘S’ is true if and only if S” (2007: 372). Again, this characterization seems to conflate truth predicates with truth concepts, and we are interested here in identifying the former rather than in offering accounts of the latter. In the remainder of the section, we will survey some distinctively alethic uses of ran in terms of the four criteria above from sections 1.1–1.4 in order to arrive at a characterization of ran that acknowledges its utility in alethic contexts without relying on theory-laden ascriptions to early Chinese thinkers.

In an illustrative example from the Analects, Kongzi uses ran to make an anaphoric, semantic endorsement of the words of one of his students, referred to simply as Yong. At the outset, Kongzi claims that someone named Zisan Bozi has a number of qualities that make him suitable for holding office. Yong replies with a modest suggestion and corrective to Kongzi’s appraisal, to which the latter simply responds, “Yong’s words are ran” (Analects 6:2).12 In doing so, Kongzi anaphorically restates and semantically endorses Yong’s words (yan 言) by means of the predicate ran, which addresses two of our four criteria. Furthermore, although he does not do so here, Kongzi could use this same formulation to express a general commitment to Yong’s words, whatever they may be: that Yong’s words are ran implies that for any <p> Yong expresses, p. Here, the scope of “Yong’s words” is limited to the conversation in the passage, but with a broad enough interpretation of Yong’s words, (e.g., those over the course of his adult life) the formulation would suffice as an example of generalization.

The Mozi, a text vehemently and self-consciously at odds with Confucianism, contains similar uses of ran where it is most naturally interpreted as a truth predicate enabling anaphoric, semantic endorsement of statements. However, in most examples of this usage Mozi is complaining that others consider his own statements to be untrue rather than utilizing ran to anaphorically endorse the claims of others.13 For example, upon claiming that society falls into disarray when political officials are not diligent in their duties, Mozi complains that “the gentlemen of the world consider my words to be not ran” (MZ 32/33–34).14 The “words” (yan 言) in question generalize over all of Mozi’s factual claims about the way the world is, which in turn offer reasons for his own preferred political reforms. Similarly, we find Mozi complaining that these same hypothetical opponents doubt his descriptions of ghosts, spirits, and ancient kings, leading off his rebuttal with, “If they consider it to be not ran…” (ruo yi wei bu ran), where “it” refers back to the description just given (MZ 63/11, 31/83, 31/77). In all such cases, ran identifies semantic evaluations by Mozi’s interlocutors, thereby engaging in anaphora, semantic endorsement (qua rejection), and generalization, three of the four features above.

Additional relevant uses appear in the Mengzi where interlocutors ask Mengzi to endorse or reject claims about a variety of topics, wherein Mengzi often responds that what was just said either is or is not ran. One notable example uses ran to not only anaphorically endorse a statement but also engage in semantic ascent. Jiao of Cao asks Mengzi: “Everybody is able to become a Yao or a Shun (i.e., a sage)—you zhu (is it so/true/the case)?” (6B/2).15 Mengzi replies, “ran”, thereby endorsing the claim that Jiao offers up for assessment without repeating it. What’s most interesting about this case is that “Everybody is able to become a Yao or a Shun”16 was a well-known Confucian slogan by this time, and while there aren’t any quotation marks in the original text, the use of you zhu at the end indicates that the slogan in question is semantically isolated and therefore being offered up for semantic evaluation. By claiming that the slogan is ran, Mengzi commits himself to its content, namely, that anyone can become a Yao or a Shun, in turn semantically descending from a claim about language to a claim about the world. Jiao could easily have asked “Can everybody become a Yao or a Shun?”, to which the answer would be either ‘yes’ or ‘no’. Instead, he asks Mengzi to semantically evaluate the slogan itself, and the proper response is therefore a semantic appraisal by means of ran.

The Xunzi also contains prominent usage of ran as a tool for rejecting statements in question by utilizing the construction, shi bu ran 是不然, or “This is not so”. This construction appears nowhere else in the early Chinese corpus (to our knowledge) but is well-suited to the context given the Xunzi’s unique style as a philosophical text, wherein Xunzi frequently attributes positions to opponents to which he then responds.17 In this way, Xunzi’s usage of ran resembles Mozi’s more than it does Mengzi’s, being employed as a tool to dispute claims by opponents. Perhaps the most famous of these examples occurs within the context of his disagreement with Mengzi over the goodness of human nature:

In this example, the demonstrative “this” is used to refer back to Mengzi’s claim, which Xunzi deems to be not ran (Fraser 2020: 121, footnote 15). Xunzi then explains why Mengzi’s claim is false, owing to his failure to distinguish between human nature and artifice.

Another use of ran as an MTP occurs in the ancient Chinese logical connective ran ze, which can be understood as “if so, then…”, “in that case…”, or simply “then…” In other words, it could be understood as “if that’s true, then…”. Early Chinese authors and speakers use this connective to facilitate inferences both in and out of dialogue contexts to identify and conditionally semantically endorse a prior claim as the antecedent of a conditional. Consider this usage in the Mengzi:

Mengzi is here essentially supposing that what Xuan said is true—that Chu would win in a war, being the larger of the two states in this hypothetical conflict—and then inferring a general principle about the strength of a state and its performance in a war based on the hypothetical truth of Xuan’s assessment.

The construction also helpfully occurs outside of dialogues, where the anaphora appears within the context of expressing a complex thought by a single thinker rather than in the context of a discussion. Consider two examples, the first from the Liji, an ancient manual on ritual conduct:

In this example, ran is used anaphorically in two separate instances as part of the antecedent of a conditional prescription about the ritual interaction between a carriage driver and rider. It functions here as an expressive device used in lieu of repeating the antecedent of the previous conditional before negating it, facilitating word economy in the text. The authors very well could have repeated the antecedent again before negating it (e.g., “If the driver’s rank is not lower than the rider’s…”) but instead, the authors use ran to refer to the original statement anaphorically before then negating it. Modern readers may not be all that interested in the particulars of this ritual, but they should nevertheless be able to see what ran is doing here as an alethic expressive device.

All these examples and many more demonstrate ran’s ability to facilitate alethic phenomena as an MTP. Furthermore, we need not posit any comprehensive theory of early Chinese language, philosophy, or psychology in order to see it, nor must we rely on strong assumptions about the fundamentality of truth to intellectual projects. Rather, we have instead identified a core set of alethic roles ran plays that can serve as grounds for further theorizing.

Ke 可 is a modal term most often translated as “acceptable” or “possible.” Its modality ranges from the sense of permission or pragmatic acceptability all the way to metaphysical possibility, logical consistency, and indeed semantic endorsement. However, most often the sense of endorsement expressed by ke is not semantic. For example, to the question “Is it acceptable (ke) for a minister to kill his ruler?” (Mengzi 2B/8) the appropriate response would be ke or bu-ke—it’s acceptable or isn’t it—where ke offers a normative endorsement of the act in question (regicide). Similarly, a famous line from the Analects tells us that “One who does not depart from their father’s Way for three years [after his death] can be (ke) called filial,” (Analects 1:11) where ke signals the appropriateness—in whatever sense of ‘appropriate’—of applying the term “filial” to someone who fits that description. In such cases, ke is used to question, reject, or endorse the acts, events, labels, or persons in question, but it does not seem to be functioning in the alethic sense that our methodology tracks.

In some cases, however, ke is used in alethic contexts to anaphorically, semantically endorse statements. Consider the following exchange between Kongzi and Laozi from the Zhuangzi anthology:

Here, Laozi uses ke to endorse all of Kongzi’s statements anaphorically. Laozi may have nodded in assent, given Kongzi a thumbs-up, or simply said “True!” or “ran!” but instead used ke to accomplish the very same thing.

In another fascinating example, ke also enables Mengzi to engage in semantic ascent and descent. In the following paragraph, Mengzi considers two claims quoted from a rival thinker, Gaozi. Mengzi claims that one claim is ke while the other is not:

In the passage, Mengzi is discussing the relationship between the heart-mind (xin 心) and the qi 氣 or “vital force” of a person. To the uninitiated, Mengzi’s claim and subsequent explanation is a bit mysterious, but this is no obstacle to our understanding of ke as an MTP. Mengzi clearly endorses one statement and rejects the other, and he uses ke to do so. Further, by reproducing each statement and appraising them directly, Mengzi invites us to engage in semantic descent: If “What cannot be ascertained by the mind, do not seek in your qi” is ke, then we shouldn’t seek in our qi what cannot be ascertained by the mind.

One final alethic use of ke appears in the Later Mohist Canons from the Mozi. Here, the Mohists are explaining their response to the sophism “All language is perverse (bei 誖)”, which they claim is itself perverse. They argue as follows:

Here, the Mohists present those who believe that all language is perverse with a dilemma. On the one hand, someone who deems “all language is perverse” to be ke is committed to that statement not being perverse, since whatever is perverse is not ke, as the Mohists stipulate. On the other hand, if they consider “all language is perverse” to be not ke, then they are not really committed to it being a “fitting” (dang 當) statement. We will say more about dang below, but the point here is just that in spite of their terseness, the Mohists are clearly using ke to express anaphoric, semantic endorsement and rejection. As this and the preceding examples show, ke indeed has the capacity to function as an MTP at least on a part-time basis within the context of early Chinese linguistic evaluations.

It would be remiss of us to fail to mention one of the most ubiquitous evaluative terms in ancient Chinese philosophy of language in this discussion: shi 是, “this” or “right”. In early Chinese philosophical writings, shi and its opposite, fei 非 (“not-this” or “wrong”), serve as fundamental evaluative attitudes by which early Chinese thinkers distinguish things from one another and establish kind membership, with shi 是 being an attitude of endorsement and fei 非 being an attitude of rejection.20 To shi something, or to deem it shi, is to accept it as a member of a kind (lei 類), being most generally rendered as “this” or “like this.” For example, Mengzi says, “The sage is the teacher of a hundred generations. Bo Yi and Hui of Liu Xia are like this (是 shi),” where shi here indicates membership of Bo Yi and Hui of Liu Xia among the sages (7B/15). Additionally, there is a more general, normative sense of shi as conduct or speech being right in an ethical sense, such as in the Mengzi where Mengzi has his conduct questioned regarding various gifts he was offered. The student, Chen Zhen, poses a dilemma to Mengzi, that if his conduct was ethically right (shi) in one case, then it must have been wrong (fei) in the other case, or vice versa (2B/3). Furthermore, the Mozi writes of Mozi rejecting (fei) the doctrine that “fate exists” (you ming 有命), on the grounds that those who believe that fate exists do great harm to all under Heaven.

Much ink has been spilled over this last example, and to a lesser extent the first and others like it, regarding the precise sense of “rejection” and “endorsement”, respectively. Why not just write for the first example, as James Legge did in his translation, that sages being teachers of a hundred generations “is true of Bo Yi and Hui of Liu Xia”? Similarly, why not just say that Mozi believes that “Fate exists” is false? This way of approaching the issue, that is, by analyzing the properties of shi and fei and identifying the sense of endorsement implied therein, has shaped the discussion between scholars such as Hansen, Van Norden, Fraser, and McLeod, about the precise sense of endorsement implied by shi and fei, specifically in the context of Mohist epistemology.21 Our chief worry with this method is that it puts the cart before the horse in terms of truth. These approaches focus on analyzing the properties of shi and fei in theoretical contexts in order to determine if that property overlaps with some contemporary conception of the property of truth, without due consideration to whether or not shi and fei ever function as truth predicates. We propose the reverse approach: first, look to everyday uses of shi and fei to see if the terms function as MTPs. Then, we can offer fuller analyses of the metaphysical or conceptual implications of those terms. This approach neither presupposes nor commits early Chinese thinkers to a substantive truth property via shi and fei. Rather, it identifies shi and fei as MTPs, and then looks to the broader use of shi and fei to help us to understand why these terms might have eventually functioned as MTPs.

As it turns out, decisive uses of shi and fei that meet the MTP criteria are uncommon but not entirely absent in the corpus. Perhaps the clearest use of shi as an MTP appears in the Analects, where yet again a student attempts to correct Kongzi, or at least question how what he just said squares with something else he said previously, to which Kongzi responds, “My disciples, Yan’s words are shi. My previous words were in jest” (Analects 17:4). This example clearly meets the standard of anaphoric, semantic endorsement. Another excellent example featuring fei also comes from the Analects, where it is featured alongside ran:

Here, the use of fei is clearly semantic because it refers to Ci’s belief, not the fact that Ci has the belief. When Ci says ran, he expresses agreement with Kongzi’s impression of his belief, saying in effect, “Yes, indeed, I do believe that.” But the fei he follows up with refers to the content of the belief, in effect asking, “is it not the case?” or “is my belief false?”, where “it” is the content of the belief. Kongzi then confirms that the content of his belief is indeed false by repeating fei and clarifying his approach to learning.

Thus, we can bypass much of the debate over whether or not all of the (perhaps several) concepts and properties associated with shi and fei overlap with contemporary notions of truth. What matters for our purposes is that because we find shi and fei at least sometimes used to express semantic, anaphoric endorsement, they sometimes function as MTPs, and so sometimes express TRUTH. It is therefore unsurprising that other properties of shi and fei may include elements we would expect to find when talking about truth properties. Nevertheless, it is important to keep in mind that these uses of shi and fei as MTPs are rare, as they are not typically applied to truth-bearers and the sense of rejection is not often exclusively semantic.

The compound term you zhi 有之 also appears in alethic contexts and appears to function as an MTP, though its usage is quite limited. The word you means in such contexts “to be” or “to exist,” while zhi is a pronoun. Combined, they typically mean something like “it happened” or “it exists,” but in a number of cases—such as in that of wondering if everybody can be a Yao or a Shun above—the most natural reading is as asking for the endorsement or rejection of the semantically isolated statement in question. In effect, it is asking: true or false? When it appears in questions like those mentioned above in the Mengzi, to which ran or bu ran22 was the answer, it appears as a contraction of a longer phrase, you zhi hu 有之乎, where hu is a question particle added on to you zhi (Pulleyblank 1995: 41). There are many examples throughout the early Chinese corpus of you zhu functioning in such alethic contexts, and quite often the response to questions containing it is a restatement of you zhi. For example, in the Zhanguo Ce, a historical text from the Han dynasty, an interlocutor says, “I have heard that your majesty has received jade and horses from the state of Wei—is it so/did it happen/is it true (you zhu)?” to which the king replies, “you zhi” (Zhanguo ce, “Qin Ce”, Section 5.8). This is clearly an anaphoric endorsement of the claim that the interlocutor has heard and brought up for endorsement or rejection. The term is not nearly as widely utilized as ke or ran—in alethic contexts or otherwise—and so there is no guarantee of examples of it being used in ways that fully satisfy all of the four conditions above, but even these limited examples in Mengzi and the Zhangguo Ce indicate its utility in alethic contexts.

Another term worth mentioning is dang 當, or “fitting.” This term has interested scholars in the debate over truth in ancient Chinese philosophy primarily because of its fascinating appearance in the Later Mohist Canons where it can be plausibly interpreted as referring to a property of correspondence between language and the world. Indeed, outside of the Canons, the term is often used to describe the “fittingness” of one’s conduct, either in general or in the context of conforming to a particular norm (Fraser 2020: 120–121). In the Mohist Canons, however, it is used to refer to the “fittingness” of predications or statements. For example, in the Mohists’ explanation of bian 辯 (“disputation”), they declare that the winner of a disputation is the one who makes the “fitting” claim.23 As we saw above, the Mohists suggest that people committed to a statement being ke should also be committed to it being dang, meaning that to endorse a claim implies a commitment to it being “fitting” in the right sort of way (perhaps “fitting the fact” as Graham (2003: 445) suggests). Additionally, the text refers to “the dang of one’s statements” (MZ A14; Fraser 2020: 120), which suggests that dang for the Mohists might be a substantive semantic property that could be developed via a theory of truth.

In spite of these fascinating possibilities, MTPs on our understanding first and foremost play expressive roles that enable speakers and writers of a natural language to engage in the alethic activities we identified in section 1. However, in the Mohist Canons, dang instead seems to be an example of a word that denotes a correspondence property between language and the world, which is precisely why it has been of such interest to scholars in this debate (who have tended to focus on the more theoretical aspects of thinking that involves truth). For example, both Fraser and McLeod end up arguing that dang does express the concept of truth on the basis of its usage in the later Mohist Canons, but neither are able to offer examples of dang functioning as a minimalist truth predicate either in the Canons or elsewhere in early Chinese writings. In other words, it seems that dang may express a candidate truth property without functioning as a truth predicate in the broader context of early Chinese language and thought.24

Our final honorable mention is cheng (誠), which usually means “sincere” or “trustworthy.” The term is featured prominently in the Confucian classics the Da Xue and Zhong Yong, where it is treated as an important virtue referring to a correspondence between one’s inner character and outward manifestations of goodness. Cheng guarantees that one’s worthy deeds are reflections of good intentions and are not merely used to disguise bad intentions (Da Xue, Ch. 3). It also acts as an adverb to emphasize authenticity, that something is “indeed”, “genuinely”, or “truly” the case, such as when Yin Shi in the Mengzi exclaims of himself that he “indeed (cheng 誠) is a petty person” (2B/12).

In some rare instances also, cheng serves as a predicate for truth-bearers that satisfies the MTP criteria. One example, noticed by Christoph Harbsmeier, appears in the Da Dai Li Ji, where Kongzi claims that a student is not fit for office. The student replies, “That I am not good enough is genuinely so (cheng ye)” (Harbsmeier 1989: 134). In the Analects, we also find Kongzi making a claim about effective government and then following up with “this saying (yan) is indeed cheng!” or, as Legge translates it, “True indeed is this saying!” (13:11). We shouldn’t be surprised to find a term that expresses both the virtue of sincerity and an adverb signalling authenticity or genuineness functioning as a truth predicate. What is more surprising is how infrequently it is used as an MTP in comparison to its far more common non-alethic uses.