Several traditions teach that the highest thing cannot be grasped in words. Their students then hold on to the words. Emptiness becomes a doctrine, and Nothing becomes a name for God. So these traditions write commentaries: a root text makes the claim, and the commentator warns the reader not to take it as one more thing to believe.
Below are four such commentaries: in Sanskrit, Classical Chinese, Latin on Hebrew, and Latin on Greek. Each is a book in this library with an English translation made by machine and not yet checked by a specialist. For each I give what the page says, what our English does with it, and where it went wrong. Quotations are our English, copied from the page they link to, and I read every quoted page against its scan. Several errors turned up while I wrote this; the ones we could fix are fixed, and the list is at the end.
1. Nāgārjuna, and Candrakīrti on the pot
Nāgārjuna’s Mūlamadhyamakakārikā (second or third century) is the founding text of the Madhyamaka, the “middle way” school of Indian Buddhism. In chapter 24 an opponent objects that if everything is empty, nothing can cause anything, and the Buddha’s teaching falls apart. Candrakīrti’s commentary, the Prasannapadā (seventh century), answers with a clay pot. If the pot existed by its own nature, it would need no clay and no potter:
Therefore, by accepting intrinsic nature, you contradict all of these things starting with the effect. In this way, if you accept intrinsic nature, nothing at all is logical. However, for us who maintain the shunyata (emptiness) of the intrinsic nature of things, all of this is possible. For what reason? Because we say:
That which is dependent origination, we call that emptiness.
That is a dependent designation; it alone is the middle way.Candrakīrti, Prasannapadā, ed. La Vallée Poussin (1903), quoting Nāgārjuna 24.18, p. 503
For Candrakīrti, things are empty because they arise from conditions: the pot depends on clay, a wheel and a potter. The verse’s second line applies the same point to the word “emptiness”, which is also “a dependent designation”. The reader is not meant to treat emptiness as a thing that exists.
The English. It keeps the Sanskrit terms (śūnyatā, svabhāva, pratītyasamutpāda) and glosses each once. Where the 1903 editor added a word in square brackets, the English marks it as an insertion. Our Sanskrit translations do worst on commentary: measured against published translations, they leave something out on 70 percent of pages, mostly by shortening the commentary (#5695). On this page the commentary is complete. On the page before it was not: the English stopped after verse 15 and left the rest of the page untranslated. That page has now been translated again.
2. The Diamond Sutra, and a Chan master named Wukong
In the Diamond Sutra the Buddha tells Subhūti that he knows the minds of all beings, because “all minds are not minds; they are called minds”. Then: “the past mind cannot be grasped. The present mind cannot be grasped. The future mind cannot be grasped.” A commentary printed in 1787 at the Haizhuang monastery in Guangzhou (金剛般若波羅蜜經如義) puts its notes above the sutra on every leaf. Above this passage it says:
Previously, all external appearances were emptied; here, the mind itself is emptied. As the saying goes, “If the skin no longer exists, where can the hair attach itself?” The writing in this chapter is ingenious and the meaning is profound, like a range of peaks where a single mountain top suddenly rises ten thousand feet high, making it difficult for people to grasp.Jingang bore boluomi jing ruyi (Guangzhou, 1787), Bavarian State Library copy, p. 51
The last clause in Chinese is 使人扳攬不及: the peak is too high for anyone to reach up and take hold of it. That is the commentator’s point. He does not offer a better account of the mind; he tells the reader there is nothing here to hold. “Difficult for people to grasp” is accurate but weaker than the Chinese.
The English, before and after. On the same leaf the sutra asks 如來有佛眼不, “Does the Tathāgata have the Buddha-eye?”, and Subhūti answers 如是世尊, “So it is, World-Honored One.” The 不 at the end of the question marks it as a question; it does not mean “no”. Our first English read it as “no”: “No, World-Honored One, the Tathagata has the Buddha-eye.” It did the same with the sand of the Ganges two lines later, and with the flesh-eye and the heavenly eye on the facing leaf. A 1496 Korean woodblock of the same sutra, also in this library, has the same sentence right: “Indeed, World-Honored One, the Tathāgata has the Buddha eye.”
We tested the two models we use. We translated the affected pages again, three times each, with the cheaper model that made the error and with the stronger one we have used for Chinese since early October. The stronger model answered “Yes” or “It is so” in every run. The cheaper one reversed the answer on the same pages every time, and in one run wrote that the Tathāgata “does not possess the Buddha-eye”. A search of every Chinese, Japanese and Korean book in the library found the same contradiction on three pages, all from the cheaper model. They have been translated again with the stronger one (#6116), and the page now reads: “It is so, World-Honored One. The Tathagata possesses the Buddha-eye.”
| Page | First English | Stronger model, 3 runs | Cheaper model, 3 runs |
|---|---|---|---|
| 1787 commentary, Buddha-eye | “No” | “Yes” 3 of 3 | “Yes” 3 of 3 |
| 1787 commentary, flesh-eye and others | “It is not so” | “Yes” 3 of 3 | reversed 3 of 3 |
| Huineng’s commentary, Dharma-eye and Buddha-eye | “It is not so” | “Yes” 3 of 3 | reversed 3 of 3 |
| Huineng’s commentary, the sand | “Yes” | “So it is” 3 of 3 | “Yes” 3 of 3 |
| 1496 Korean woodblock, Buddha-eye and sand | “Indeed” | “Yes” 3 of 3 | Buddha-eye answer lost its “yes”, 3 of 3 |
Chan turned the same teaching into dialogue. Xiufu Wukong was a tenth-century master at the Qingliang monastery, which the ruler of the Southern Tang built and invited him to lead. His name, 悟空, means “awakened to emptiness”. His exchanges with monks are recorded in the Compendium of the Five Lamps (五燈會元), compiled in 1252; ours is the Qing-dynasty copy made for the Siku Quanshu imperial library. A monk asks, “What is the Way?” He answers with the best-known line in Chan, from the verse attributed to Huineng:
Originally there is not a single thing; where could dust collect? The monk bowed. The Master said: Do not misunderstand. A monk asked: What is meant by “one speck of dust entering into right concentration”? The Master said: Form is emptiness. The monk asked: What is meant by “arising from the samadhi of all dust particles”? The Master said: Emptiness is form. The monk asked: I will not ask about other matters, but what is the single phrase for realizing emptiness? The Master said: It is two phrases.Compendium of the Five Lamps, vol. 8, the monk’s question on scan 152, p. 153
The monk bows as if he has been given a doctrine, and Wukong tells him not to misunderstand. Asked for the single phrase for realizing emptiness, he answers that it is two: the Heart Sutra’s “form is emptiness” and “emptiness is form”, and neither can be dropped. The question also contains his own name, 悟空, so the monk is asking for “Wukong’s one phrase” as well.
The English, before and after. Our first English kept the name and put its meaning in a note, so a reader could see both senses. The new English gives only “realizing emptiness”, and the name is lost. It was translated again because the first version had a worse error a few lines further down. The monk asks, “What is my own fundamental affair?” and Wukong answers, “How many places have you gone to ask people this?” The monk then asks what the ancients attained that let them rest, and Wukong turns the question back on him. The first English ran the answer into the next question and gave both to the master:
The Master said, “How many places have you asked people about what the ancients obtained before they stopped and rested?” The Master said, “What have you obtained that makes you unable to rest?”
It also glossed Niutou, the monk Niutou Farong, as “Ox-head mountain”. The new English has the right speakers and the right Niutou. Wrong speakers are a common error in our translations of argument and dialogue. In a blind review of 150 random pages of our Tibetan Tengyur translation, about one page in four had a reversed statement or the wrong speaker or agent, most often in logic and in verse (#5829). Across our measured Chinese pages, 18 in 100 had a reversal of some kind (#5695).
3. Ayin: a Latin dictionary of Kabbalah
Christian Knorr von Rosenroth’s Kabbala denudata (Sulzbach, 1677) opens with a dictionary of kabbalistic terms in Latin, drawn from Moses Cordovero’s Pardes Rimmonim, Joseph Gikatilla’s Gates of Light and the Zohar. For a long time it was the fullest account of Kabbalah in Latin. Its entry for Ayin, “nothing”, includes:
All agree that Ayin is a title of the Crown; and this is because no one can grasp it; but if a question is raised concerning it, the answer must be: that which is perceived or understood of it is Ayin. Rabbi Moses, however, provides another reason: namely, that a “Principle” is usually called “Non-being”; and this is the Principle of all Principles.Knorr von Rosenroth, Kabbala denudata (1677), Bibliotheca Philosophica Hermetica copy, Amsterdam, p. 81
Here Ayin is the name of the Crown, Kether, the highest of the ten sefirot. It is called “nothing” because whatever is understood of it is not it. On the page before, Knorr notes that אין, ayin, “nothing”, and אני, ani, “I”, are written with the same three letters. Ayin stands for the first sefirah and Ani for the last, and blessing flows down through the channels from the one to the other.
The English. It keeps Ayin, Kether, En-Soph and Malchuth as names, gives the Hebrew, and puts Knorr’s Latin Nihil in a note. It also adds a guess in a note: that “Rabbi Moses” is “likely Moses Cordovero”. The guess is reasonable, since the entry cites Cordovero’s Pardes twice, but it is ours and not Knorr’s. In our reference study, Kabbalah and mysticism scored lowest among the Hebrew, Arabic and Persian texts, 3.04 out of 5 against 4.00 for scripture and law (#5695), mostly because of misread Hebrew type. This page is in Latin and reads cleanly. The scans of this book have a different problem: printed pages 80 and 81 were photographed twice, so a reader turning the pages sees them in the order 80, 81, 80, 81.
4. Dionysius, on a page that compares translations
The fifth chapter of the Mystical Theology of Dionysius the Areopagite (about 500) is the standard statement of negative theology in Greek Christianity. The cause of all things is not soul, mind, number, unity, goodness or spirit, “nor darkness, nor light, nor error, nor truth”. Johann Eck’s edition of 1519, a copy of which belongs to the Embassy of the Free Mind, prints the Greek beside three Latin translations: by Johannes Sarracenus (twelfth century), Ambrogio Traversari and Marsilio Ficino (both fifteenth). Below them are a paraphrase by Thomas Gallus, abbot of Vercelli, and Eck’s own commentary. Eck sums up the chapter:
Finally, he takes away all beings and non-beings in general. He adds that He is not understood by anyone, nor does He understand anything. Finally, he concludes that there is no speech or knowledge of God, because He is above every assertion and negation, and simply absolute from all, and more eminent than all.Johann Eck, commentary on Dionysius, De mystica theologia (1519), Embassy of the Free Mind, p. 66
God is above negation as well as assertion, so “God is not” is no more an answer than “God is”. Like Candrakīrti, Dionysius applies his denial to the denial itself.
The transcription. The Greek column ends by saying the cause is “beyond” something. In the standard text it is ἐπέκεινα τῶν ὅλων, “beyond the whole”, and all three Latin versions on the page agree: super tota, totisque superior, ultra omnia. The Greek printed in this book reads ἐπέκεινα τῶν λόγων, “beyond words”. Our transcription gives τῶν ὅλων, the standard reading, not the one on the page. Whether λόγων is a variant or a printer’s error, a reader comparing the four columns should be able to see it. We have found the same tendency before: a transcription of a Vatican Zohar manuscript contained well-known passages that are not on the leaf.
The facing page, where the chapter begins, has a larger error. The printed page has four columns, each with a heading: the Greek, then Sarracenus, Ficino and Traversari. Our transcription has three. It moved each heading one column to the left, so the Greek is labelled “Sarracenus”, and it left out Ficino’s column, so Sarracenus’s Latin appears under Ficino’s name. The English translates this rearranged text accurately, so nothing in it warns the reader.
Greek
his Latin
Ficino’s Latin
Traversari’s Latin
Greek
Sarracenus’s Latin
Traversari’s Latin
What the four share
The four do not teach the same thing. Nāgārjuna’s emptiness means that things which arise from conditions have no nature of their own, and there is no creator behind them. Ayin is the highest aspect of a creating God. Dionysius’ cause is beyond being and non-being. Wukong refuses to let any of it become a formula. Reading them together does not show a single teaching under different names, and an English that made them sound alike would be wrong.
They do share a method. Each states a negation and then qualifies the statement. Candrakīrti calls emptiness “a dependent designation”. Wukong says “Do not misunderstand” and “It is two phrases”. Knorr says that what is understood of the Crown is Nothing. Eck says that God is above negation as well as assertion. In each case the commentary is there to correct the reader.
Four commentaries, root text to the copy in this library
Tap or hover a mark. A band is a date the article gives only as a century.
Nāgārjuna and Candrakīrti · Sanskrit
Diamond Sutra and Chan · Classical Chinese
Knorr von Rosenroth on Ayin · Latin on Hebrew
Dionysius and Eck · Latin on Greek
7th century · Nāgārjuna and Candrakīrti
Candrakīrti, Prasannapadā, the commentary.
- The negation
- Things that arise from conditions have no nature of their own.
- Then the commentary qualifies it
- Emptiness too is “a dependent designation”.
What this means for translation
Keep the terms. If śūnyatā, ayin and Dionysius’ “beyond” all become “nothingness” in English, any resemblance a reader sees comes from the translator. Our translations keep the terms and gloss them once. In our comparison with published translations in fourteen languages, our English scored higher on keeping terms and showing its sources, and lower on readability and voice (#5695). The loss of “Wukong” in the new translation of the Five Lamps page shows that this can still slip.
Translate the page in front of you. A transcription can drift toward the text it already knows, as with τῶν ὅλων for τῶν λόγων, or lose one of four columns. Any translation made from it inherits the error, so the transcription has to be checked against the scan.
Check who is speaking, and whether the answer was yes. These texts argue in dialogue: objection and reply, monk and master, question and answer. The worst errors we found were not wrong words. They were a sentence given to the wrong speaker and a question marker read as “no”.
Translate the commentary in full. It is the part our translations most often shorten, and in these traditions it carries the teaching.
Fixed while writing this, and still open
Fixed: the reversed answers on three Diamond Sutra pages, the wrong speakers and the “Ox-head mountain” on the Five Lamps page, and the untranslated half of a Prasannapadā page, all translated again. Still open: the standard Greek reading and the missing Ficino column in Eck’s edition, a line the transcription skipped on the facing leaf of the 1787 commentary, and the doubled pages in Knorr. These need new transcriptions or new scans, and are tracked in #6116.
All of these translations are unreviewed drafts, and each page says so. If you read Sanskrit, Classical Chinese, Hebrew, Greek or Latin and find an error, the feedback link on every page reaches us.
