I am most indebted to my colleagues Ernest Thrimbe and Arlo Griffiths for commenting on this article and offering editorial suggestions.
1The Sulalat al-Salatin is the object of constant debate in Malaysia today. Conferences and lectures are frequent, publications follow each other, editions of the text keep appearing. During the last twenty years six new editions have been published in Kuala Lumpur and Malacca. This exceptional academic activity (second only to the debate, even more lively, around Hang Tuah, at the expense of all other texts, which tend to be neglected) does not produce a better knowledge of the text, but a political reading that, beyond quarrels of interpretation, rests on a patriotic celebration. It seems that interest for ancient Malay literature has drastically diminished in Malay society during the last thirty years. Today the Sulalat al-Salatin is the domain of academics—male ones exclusively—some of whom know the text by heart, so to speak, but approach it with a questioning different from that of their foreign colleagues.
2Sulalat al-Salatin is the title inscribed in the text itself: it is the title chosen by the author. However, the text is more frequently quoted with the title Sejarah Melayu, translated into English as Malay Annals.
3The prestigious series Karya Agung, which had already published an edition of the text (by Muhammad Haji Salleh) in 1997, recently published a second edition by Dr. Ahmat Adam (see Ahmat 2016 in the Bibliography below). Ahmat Adam’s edition is the eighteenth edition of the Sulalat al-Salatin published since the middle of the 19th century.
4Ahmat Adam (henceforth A.A.) gives an edition of the “Krusenstern manuscript” (henceforth K), i.e. a manuscript acquired by Commodore Krusenstern in Malacca in 1798. We need to consider that date for a moment. Ivan Feodorovich Krusenstern was a Russian of German origin, also known as Adam Johann von Krusenstern. Before being appointed captain in the Russian imperial navy he served for a while on commercial vessels of the British East India Company. This is when, on a travel from India to China, while his ship was under repair in Penang, he spent a few months in Malacca, in the second half of 1798, and took the opportunity to order a copy of the Sulalat al-Salatin (A.A., p. xcvi).
5The Sulalat al-Salatin has been mentioned by Nuruddin al-Raniri, a famous Malay author, in Aceh, around 1640 and by several European authors starting in 1708 (P. van der Vorm, F. Valentijn, etc.). Its contents, however, remained unknown until the publication, in 1821, of John Leyden’s translation. This is why Krusenstern’s decision to obtain a copy, while he did not know the Malay world and was spending a few months in Malacca by mere chance, is surprising. Krusenstern was probably lucky enough to be advised on the spot by some connoisseur. When back in Russia he presented the manuscript to the Royal Academy of Sciences in St-Petersburg through his friend Fyodor Ivanovich Schubert, who was a member of the Academy and gave the manuscript to it in 1802 (Kulikova, p. 28-29). The manuscript was copied in 1213 A.H., i.e. June 1798 to May 1799. As Krusenstern was already in Canton in November 1798, the manuscript has to date from June to October 1798 (Kulikova). It bears a watermark dated 1794, which fits perfectly with that estimate. This makes K the oldest copy we know of the Sulalat al-Salatin, albeit by a few years only: the following oldest manuscript dates from 1808.
6Manuscript K is still kept in St Petersburg. It has been described by A.M. Kulikova in 1989 and subsequently published for the first time by E. Revunenkova in 2008. That edition consists in a facsimile (with no transcription) and a 105-page commentary in Russian. This work has remained unknown outside Russia because of its language. A.A., who does not know Russian but obtained an insight into the contents of the book thanks to a Russian colleague, states that Revunenkova’s reading and analysis are deficient; she can’t read the Jawi script properly (p. xc). I don’t read Russian either but E. Revunenkova has published an article in English in 2006 (listed in A.A.’s bibliography but nowhere commented upon) that would rather indicate that she has read and studied the text in a careful and competent way.
7A.A. reproduces the facsimile in his own book (as the result of a recent and excellent policy of the Karya Agung series). A.A.’s book therefore consists in a long introduction (120 pages), the annotated edition of the text, and the facsimile of the manuscript. I suppose nobody in the world, even in Malaysia, will choose to read the Jawi text in its original form rather than the Latin transcription any more, but the publication of a facsimile is of great benefit for philologists; it may sometimes also be a devastating element, similarly to bilingual editions, when a punctilious reader sets out to compare facsimile and transcription.
8The K manuscript is in excellent condition and the facsimile is of a good quality, even if some pages are sharper than others. A.A. worked on the facsimile without any access to the original. He reproduces the description of that original by A.M. Kulikova (1989), without noticing that, in that description, the “first” pages designate the last and vice versa. This apparent mistake is due to the fact that Kulikova refers to the pagination, added by a librarian and which has disappeared from the facsimile, ordered from left to right, like in a European book, and not from right to left, as it is the norm for manuscripts written in Arabic characters (this is explained by Revunenkova 2006: 63). Therefore that kind of information in A.A.’s book (p. xcii) has to be mentally converted.
9In his presentation of the facsimile, A.A. follows three uncommon principles: a) he numbers the folios in Roman numerals; b) he starts with folio ii; c) he considers a folio as made of two facing pages (a double spread). He fortunately preserved the original presentation of the manuscript, the text starting on a verso, the layout of which, together with the following recto, makes a kind of frontispiece. Therefore, his facsimile starts with a verso called f ii, followed by a recto called f ii:2. This is somewhat confusing but has no consequence on the reading or commenting of the text. In the following notes I convert the numbering into Arabic numerals: f lxxxviii and f lxxxviii: 2 become f. 88:1 and f. 88:2 respectively. Furthermore, the facsimile includes one double spread printed twice (the “folio” 400), while the preceding double spread is lacking. One more page is printed twice too—an issue to which we will return below.
10The original manuscript is made of two volumes. On the last page of the first is inscribed the date 1213 and an Arabic sentence according to which, in A.A.’s translation (p. xcv), the manuscript has been copied by three men (Al-Haj Muhammad Tahir al-Jawi, Muhammad Zakat Long and Ibrahim Jamrut), who have been paid by the day. A.A. asserts that the name (nisba) al-Jawi means that the first is of Javanese origin (p. xcvi), whereas it means from Southeast Asia or more specifically from Sumatra. Ibrahim Jamrut would be of Javanese origin too because his name is the Javanese prononciation of the Malay word jamrud (p. xcviii).
11A.A. makes no comment on the fact that the copy was made by three scribes. A cursory examination of the facsimile makes me think that the handwritings of three different people (say, K1, K2, K3) can indeed be distinguished: K1 is responsible for two passages: ff. 2:1 – 10:1 and ff. 38:2 ‒ 105:2, that is, the first 17 and last 135 pages of the first volume, with a total of 152 pages; K2 has copied ff. 10:2 ‒ 37:2, that is, 55 pages in the middle of the first volume; K3 has copied the totality of the second volume (ff. 107:1 ‒ 192:1), that is, 171 pages. These conclusions are provisional, but they raise interesting questions. When he reaches the last two pages of his first section (the beginning of the text) K1 enlarges his handwriting considerably, in order not to create a break with the following section, which is being written by his colleague K2. The latter too, at the end of his own section, enlarges his handwriting, but cannot avoid a blank page between his section and the following one (f. 38:1). There is one more blank page further on (f. 142:2), but that does not seem to be related to a change in handwriting.
12It seems thus clear that the three scribes worked simultaneaously. The fact that they shared the task in that way may indicate that they were in a great hurry, which might explain the mediocrity of their work. What models did they use? Four fragments of an unbound manuscript or several manuscripts? The use of several manuscripts could explain that the final text is difficult to classify among the different versions of the Sulalat al-Salatin. It would be useful to scrutinize more thoroughly the passages written by the respective scribes because that could perhaps explain spelling inconsistencies and some idiosyncrasies (for instance a few typically Javanese spellings or the alternating final b/p); moreover, it is probable that the quality of the copying, mediocre as a whole, is not the same with one copyist and another.
13A.A. asserts a few times that K is a copy of good quality, much better than that of the most famous manuscript of the text, Raffles 18 of the Royal Asiatic Society (henceforth manuscript R18), but the unique criterion of that excellence lies in the claim that Old Javanese words are better preserved (p. xciv, cxix). But A.A. also notes that K is “careless and negligent,” as can be seen from spelling inconsistencies and mistakes, repetitions, and lacunae (p. xcvii-xcviii, also p. xxx fll.). A.A. notes a significant number of copying errors, and one discovers more and more of them while reading the text. On the whole, K is a rather careless copy on the literal level (both misreadings and miswritings); on the other hand, one gets the feeling that the scribes followed their models faithfully (in a hurry they had no leisure for improvisation). The consequence of these remarks is that it is not possible to rest the authority of a reading on the spelling of the manuscript (as A.A. frequently does) and that in an edition that claims to be “critical,” it is necessary to amend the text wherever it is obviously incorrect.
14It may be useful to say a few words about some peculiarities of the manuscript’s orthography that have a decisive influence on the reading and the transcription of the text: 1) the dal is often written like a lam (p. xlvi); 2) two words (Rabingul-awal and saringat, perhaps others) are written with a nga (ڠ) instead of a ‘ayn, which is typically Javanese; 3) the writing of a final <b> for a <p> (voiced for unvoiced) is frequent (e.g. adab for adap, atab for atap, bercakab, genab, hidub, berlengkab, etc.), while the reverse (<p> for <b>) appears only once (takjup for takjub); 4) it is often difficult to distinguish one or two diacritical points; 5) the name Allah, of very frequent occurrence, is systematically written with two <l> so small that they look like one <s>, so that the word seems to be written <ash>, with a <h> reduced to a minuscule slanting stroke; so much so that the same lettering is once transcribed <asih> (p. 14 l. 2).
15The Karya Agung series is one among many, worldwide, that aim at publishing a “canon,” the major works, of a literature. It is printed with a paper colour and ornementation inspired from the famous “yellow books” (kitab kuning) used in Koranic schools, but in a luxurious (and expensive) fashion, like rare and precious objects that one is proud to exhibit in one’s personal library. However, since a few years back, each volume is also printed with a soft cover at a much more reasonable price, which allows this series to be the principal publisher of Malay classical texts today. The series’ editorial board claims to be extremely rigorous philologically and the recent decision to publish a facsimile of a manuscript in each edition does reinforce this “scientific” aspect. The targeted audience is Malay society at large but some editions, because of their imposing critical apparatus, rather look like academic exercises.
16A.A.’s edition of the Sulalat al-Salatin definitely belongs to that category. The transcription is congested with a myriad of superfluous signs and annotations: a) in the manuscript the letter sin is used to transcribe the phonemes /s/ and /š/ (<sy>), instead of /s/ alone; A.A. transcribes all the words comprising a /š/ written as a sin as <s[y]>; we thus have <S[y]ah> a hundred times. This is an unnecessary hyper-correction: Jawi is not Arabic; sin transcribes <sy> in a perfectly standard way in many Malay manuscripts; b) A.A. introduces into the text, inside square brackets, the Jawi lettering of difficult words, despite the fact that the whole Jawi text is published in the same book; c) an enormous amount of archaic spellings are followed by “[sic]” (“menengar [sic], tuha [sic], tahta [sic], ra’na [sic], nentiasa [sic],” etc.), so that sic must be one of the most frequent words of this edition, beside maka, pun, yang or telah; d) the text is accompanied by 1,579 footnotes, the majority of which quote P.J. Zoetmulder’s Old Javanese Dictionary. All this gives the edition an aspect of high learning, but makes it rather indigestible and confusing.
17Malay philology is at a critical point of its history, because there is presently no debate on theories and methods, no handbook that would offer more than general considerations, and no individual reflexion that one would be able to find in the edition of a particular text. Even if there have been some debates or pronouncements since then, the sole theoretical synthesis on the subject is a 15-page article by a German scholar published 36 years ago (see Kratz, 1981). In these conditions, each “philologist” deals with “his” text with common sense as only guide.
18Let’s see some of the choices made by A.A., considering that he has decided to publish a critical edition, as stated in the very title of the book (disunting dengan kritis). There will be plenty of opportunities to observe that “critical edition” here means a transcription of a manuscript corrected by the editor according to his taste, his knowledge and his hypotheses, most often without any consideration for the readings of other editions, other than to declare them faulty.
19The first of these choices is the division of the text into chapters: the text of the Sulalat al-Salatin is divided into sections that begin with the formula “Alkisah maka tersebutlah perkataan” (“Here now is the story of”) and end with the formula “wa’llahu a‘lam” (“God knoweth the truth”) in a more or less elaborated form. All editors have regarded these sections as chapters and have given them numbers. So does A.A. “demi memudahkan pembacaan” (p. xciii). It is not necessary to number chapters to “facilitate reading”. Editors of ancient Malay texts are rarely conscious of the way they manipulate a text even though they claim to reproduce it faithfully.
20A second choice regards punctuation: Malay Jawi texts have none; all editors (rightly) add one to their transcription in Latin characters and most tend to create short sentences, particularly by introducing a point before every occurrence of the word maka. This is more important than it seems because punctuation structures the text, imposes a reading rhythm and determines an interpretation. A.A. shares the proclivity toward short sentences and it happens that his punctuation hinders the reading instead of guiding it. It can even create misinterpretations. (In order to spare the readers I will quote two examples only, among many others, in each category below.) So, two examples of faulty punctuation: a) p. 9 l. 3, ‘“raja Iskandar anak raja Darab Rum, bangsa Makaduniah nama negerinya” is a mistake for “Raja Iskandar, anak Raja Darab, Rum bangsa[nya], Makaduniah nama negerinya”; b) p. 322: “Maka kata orang Melaka, ‘Ini Benggali putih’ pada seorang Feringgi itu. Berpuluh-puluh orang Melaka mengharu dia.”, whereas one should read: “Maka kata orang Melaka, ‘Ini Benggali putih’. Pada seorang Feringgi itu berpuluh-puluh orang Melaka mengharu dia.”, which is found in all other editions.
21In the third place—but this is certainly not a matter of choice—one finds a certain amount of mistakes in the transcription. Two examples: a) p. 10 l. 8, “jadi Islam di dalam ukum [hukum] Nabi Ibrahim” is a mere blunder: the lettering <akm> cannot signify [h]ukum and should obviously be read as agama; b) p. 38 l. 12-13, the transcription (twice) panjar-panjar (followed by “[sic]”) is faulty: the manuscript has panja upanjara. A reading mistake of a different kind is found on p. 40: A.A. asserts (fn. 353) that a whole page of the manuscript has been copied twice and he comments on the scribe’s sloppiness, but in fact, it is not the manuscript that repeats itself, it is merely the facsimile (the comparison of ff. 19:2 and 20:2 leaves no doubt). This means that the facsimile has been established page by page and that, starting on this point, versos have become rectos and vice versa: the whole manuscript and the pagination are from here onward disorganised.
22Fourth, as this edition is light-years away from a diplomatic one, evident mistakes—and there are plenty—should have been corrected in one way or another. Two examples from the same page: a) p. 33, in “maka baginda pun naik ke atas keinderaan baginda lembu putih,” it is clear that keinderaan is a mistake for kendaraan; b) p. 33 l. 17, kembalilah; the word has no meaning here and should be corrected to kayalah (the reading of all other published versions). Not to correct these mistakes, either in the text or in footnotes, will inevitably confuse the reader.
23Fifth, and this is the reverse flaw, some corrections are excessive. Two examples: a) the phrase “Nusyirwan Adil, raja Sarib Maghrib” crops up several times (e.g. p. 16 twice, p. 31, etc.), and it is each time corrected into “raja masyrik-maghrib,” which is the reading of Abdullah’s edition as well as R18; this would have been justifiable once but not several times; considering the alternate letterings s/sy and b/p, I think syarif (descendant of the Prophet) is meant, that is, “king of the syarif of the West”; b) p. 26, l. 22, the text has “hikayat Hamzah” (like Abdullah and R18), with a slight spelling error; A.A. corrects into “hikayat Hamurabi”; the Sulalat al-Salatin is not an erudite text; its authors did not know much about the outside world and even less about history, while copying mistakes are innumerable. Common sense dictates that the hikayat mentioned here is the famous Hikayat Amir Hamzah, which is mentioned in the same way (“hikayat Hamzah”) elsewhere in this same text (A.A. p. 351), not a fanciful hikayat Hamurabi.
24Sixth, A.A. introduces into his text, which he claims to represent the most ancient version of the Sulalat al-Salatin, passages borrowed from other, more recent, versions of the text (e.g. pp. 24, 39, 184, 337), in order to fill lacunae. It is of course necessary to signal lacunae and to summarize their content, but to integrate into a version several pages of another is to bring about voluntary contamination.
25The seventh choice represents one of the major characteristics of this edition; it concerns the allegedly Old Javanese words. A.A. has the theory that many words in the Sulalat al-Salatin originate from Old Javanese and that they are particularly numerous in manuscript K, which in turn would prove the antiquity of this version. Winstedt once published a short note on “Sanskrit in Malay Literature” (1957), in which he asserts that in classical Malay texts, and particularly the Sulalat al-Salatin, words of Sanskrit origin are four times more numerous than those of Arabic origin. His evidence is scanty but the idea is suggestive. A.A. has been struck by the same phenomenon, but he talks mainly about words of Old Javanese origin, and this amounts to confusing several things: the fact that a word exists in both Malay and Old Javanese does not mean that it has been borrowed by the first from the latter; it may have been borrowed independently by both languages from Sanskrit (or from another language: A.A. also claims that the word jonk [junk], well-known to be a loanword from Chinese, is of Old Javanese origin, fn. 439). Moreover, even if it were established that a Malay word is of Sanskrit or Old Javanese origin, it would not necessarily have a spelling identical to that of the original word (not to mention the possibility of spelling variation in the source language itself). What is more, a word of foreign origin does not always have the same meaning in the source and the target languages. Still, in blatant contradiction to all such well-known facts, A.A. wants his readers to believe that “It is only when the origin of each word has been examined etymologically that its correct form can be known” (p. xxxiii).
26A.A. draws conclusions from the quantity of those “Old Javanese” words: the author would have borrowed them directly from ‘classical’ Javanese texts (p. lii), therefore the author of the first version of the Sulalat al-Salatin, that at the origin of all others, would have been a man of Javanese ascendency, or a Malay-Javanese mestizo, in any case a man who spoke fluent Javanese or was accustomed to utilizing Old, Middle or Modern Javanese words in writing (p. xl); and he would have been an expert in Old Javanese language and literature (p. li). All this is pure fancy.
27On the basis of these convictions, A.A. comments upon all the words supposed to be of Sanskrit or Old Javanese origin. The first footnote to the text reveals that the Malay keras is borrowed from Old Javanese; elsewhere, we find notes on the meaning, in Sanskrit or Old Javanese, of the words adu, beta, citra, demang, duli, empu, mutia, niscaya, pandai, etc., etc. This is not only out of place (why not a commentary on words of Arabic or Chinese origin?) but highly ambiguous: it suggests that those words in the Sulalat al-Salatin have the meaning of their equivalent in Monier-Williams’ or Zoetmulder’s dictionaries. A.A. is proud to have “preserved” so-called “classical” spellings (p. xxxii) like karunya (for kurnia), nityasa (for senantiasa), prastawa (peristiwa), pramuka (permuka), karana (kerana/karena), and we also stumble in the text across tepramanai (tepermanai), sambrani (semberani), pramadani (permadani) and more. This does not mark any progress in philology, it is simply barbaric. “Sekali prastawa” sounds like a joke.
28This obsession with Old Javanese and Sanskrit leads A.A. to correct, often erroneously, the text of the manuscript according to Sanskrit or Old Javanese vocabulary. Two examples: a) pp. 12-13, in the phrase “dan segala ulama dan hukama meayarkan emas dan perak,” meayarkan (an incongruous creation on an Old Javanese base) is a faulty correction; the word should have been corrected to menaburkan; b) p. 21 l. 27, the lettering <a-w-ŋ> is corrected to the Old Javanese wwang (idem on p. 64), which has nothing to do here, while it is also (and correctly) corrected to o[r]ang in the very same line.
29This Old Javanese frenzy affects names too: p. 33 last line, Bota: this name has been read Bat or Bath by everybody to this day; the spelling of the manuscript is not as clear as A.A. says in fn. 324: the diacritical sign over the ta is not a fatha (sign of the vocalisation “a”) but a shadda (sign of gemination, which is perfectly clear on the facsimile, f. 20: 2, line 2), and there is no reason to “correct” to Bota. Incidentally, the lettering of the word with a shadda, i.e. <bṭṭ>, seems to confirm van Ronkel’s hypothesis (1921: 175) that the name Bat comes from the Sanskrit bhaṭṭa, “the erudite, the bard,” a typical surname for a Brahmin.
30The fascination for Old Javanese perverts a famous passage of the Sulalat al-Salatin, the ciri. It is a text of a few lines that a character of supernatural origin (Bat precisely) delivers during the consecration of a king. The ciri seems to have first been a eulogy, in Sanskrit, of the king to be enthroned, that was read by a priest. Then the text became corrupt across the ages, while its use was extended to high officers of the kingdom. From the point of view of the disparity of the versions of the Sulalat al-Salatin, the ciri is an interesting point of comparison: it is mentioned in all versions, but its text is not quoted in Winstedt’s and Abdullah’s versions; it is quoted once in recension III, and quoted twice in K and another witness of recension II. A.A. thus transcribes the ciri twice and devotes an appendix to it (pp. 373-7), in which every word is commented upon. I am too incompetent myself to discuss these transcriptions and their translation, but it seems exceedingly surreal to try and read, and translate, a text in an unknown language by thumbing up a dictionary, even more so if the dictionary is one for Old Javanese, while the language is believed to be Sanskrit. Several scholars have studied the ciri in the past (among others, R.O. Winstedt, W.E. Maxwell, Ph.S. van Ronkel), but AA feels entitled to ignore their work altogether.
31A.A.’s voluminous introduction discusses the various questions evoked above and also addresses the genesis of the text. Numerous hypotheses, all of them quite fragile I would think, have been proposed (mainly by R.O. Winstedt, R. Roolvink, O.W. Wolters, Teuku Iskandar and V.I. Braginsky) on the successive stages of the redaction of the Sulalat al-Salatin in the course of ages. A.A. adds a stone to this fanciful building: for him, the text has been revised under five Malacca sultans (p. lxiv, lxxvi)—“revised” because a first draft already existed in the 14th century (p. xlvi) or even the 13th (p. xlv, lxxiv). A.A. talks many times of a “standard” text, defined in various contradictory ways and which finally transpires to be Abdullah’s edition of 1841. This point of view was current in the 19th century but has long lost all reason to persist. A.A. still has a few more theories on the date when it was decided to draw up the 1612 version of the Sulalat al-Salatin, on the personality of the author (a Sufi of Shia tendency…), on the influence of Aceh on the text, and other questionable points of view.
32Summing up, this edition does not make the text of manuscript K easy to read. A.A.’s book is obviously a work of love and dedication. Unfortunately, A.A. got carried away with extreme ideas about philology, while a basic and modest transcription would have been much more profitable. Still, we do have a text and it transpires that K is a very interesting version of the Sulalat al-Salatin. There are three recensions of the text (plus a few hybrid versions and editions), that are commonly known as the “Winstedt,” the “short,” and the “long” recensions. Roolvink (1970: xxii) and Revunenkova (2006: 63) have stated that K belongs to the short recension, but A.A., defying all evidence, asserts, on the basis of only few and shaky arguments, that K belongs to the Winstedt recension: K and manuscript Raffles 18 would have been copied on almost identical models (p. cvi and others) and K would be the most faithful witness of the 1612 version. This thesis is a priori attractive because the “Winstedt” recension is only known through one complete manuscript and another containing half of the text only. But, in reality, a comparison of the available versions of the text shows without any doubt that K belongs to another recension, namely the “short” one, and it represents an original version of that recension by comparison with the two versions known until now. (I intend to publish the results of that comparison in another article.) K is close to Abdullah’s text in all major criteria of classification, but it is also close to either one or the other of the two other recensions in minor criteria, and yet shows idiosyncrasies of its own. Despite all its foibles, this edition is therefore extremely useful. It will require a thorough study of the text to determine the place of K in a stemma (still to be built) of all versions of the Sulalat al-Salatin, but that is another story.