WHAT WAS IT THE EDITOR THOUGHT THE TRANSLATOR TRANSLATED?

In 1692-95 the Dutch politician and ethnographer (&c. &c.!) Nicolaas Witsen (1641-1717)^{1} published his famous work "Noord en Oost Tartarye" ("Northern and Eastern Tartary") about lands and peoples East of Muscovy.^{2} In the second edition (1705) we find word lists for a number of then very exotic languages, among which Kalmuk, Dagur, Mordwinian and Korean. (For some peoples, this is the earliest document about their language!) The longest list is Georgian. Wim Lucassen is about to publish an edition of Witsen's Dutch-Georgian word list, in which he identifies the Georgian words, and compares the spelling and meanings given by Witsen with the modern ones.^{3} (This article was provoked by a preliminary report Wim Lucassen published recently.)^{4}

Such identification, on first sight, looks much easier than it is. Both Dutch and Georgian have changed since the end of the 17th century. The Georgian words in Witsen's list appear in a Dutch transcription that is at least inaccurate in distinguishing Georgian sounds. There is one translation per word (that alone is asking for trouble). The informant(s) (unknown to us) may have used regional word forms. They clarified the meaning of the words in a way unknown to us, and in a language (or languages) unknown to us, to a person (or persons) unknown to us, whose native language is equally unknown to us.^{5} Those translations have been translated into Dutch (by Witsen or someone else). In every stage of the process, errors, mistakes and misunderstandings are bound to have happened.

Wim Lucassen writes that he has managed to match nearly all 904 words in Witsen's list with modern or old Georgian words - i.e. he found an explanation for differences in spelling and/or for differences between Witsen's translation and the meaning of the word. For 25 words, in sound close or equal, however, he could not until then establish a match. Those word pairs I tried to tackle. (I have not seen the 879 others.^{6})

In those 25 cases, Witsen's transcription coincides quite nicely with one or two modern Georgian words but it is hard to see how his translation and the known meaning of the word can be reconciled. Take Witsen's "magali = a flea" and Georgian "maghali = high" or "maqali = chafing-dish". Then one would question the 17th-century translation and/or transcription - but first one should explore whether "flea" and "high", or "flea" and "chafing-dish" could be matched!

Not knowing Georgian...

The truth is, I do not know Georgian - even my knowledge about Georgian is near to none... So, how do I dare to get involved in this? My excuse would be that linguistic puzzles are not purely linked to a particular language. In this case, knowledge of Dutch and of some possible intermediate languages is of more use - as is some experience with this kind of puzzles.

During the last years, I have been involved in similar problems emerging from the study of Russian translations from mainly Dutch and German 17th-century newspapers - handwritten translations, to be read to the tsar and his boyars exclusively. I must concede that my knowledge of the Russian language of that period is very limited, too. So I am very glad that my co-author for publications on the subject, is a specialist on it.^{7} In this case, I shall be happy to leave problems about Georgian to specialists on Kartvelian.

The questions we ask ourselves in regard to those newspaper translations may be different from the problems encountered in regard to Witsen's Georgian list - the ways to find an answer are quite similar. We want to know what was the language of the original, and whether it could have influenced the language of the translation, in grammar or in the choice or the form of words.^{8} When an original newspaper cannot be found (very few 17th-century newspaper issues have been preserved^{9}), we have to try to find out what the original text could have been. When we did find a newspaper issue with an article that could have been the original for the translation, we have to check accurately whether it was indeed the source (because similar news items tend to appear in different newspapers that may have all been lost).^{10} When there is a hole in the only surviving copy, then, too, we have to imagine what it was the translator translated (and check whether our findings fit in that hole!).^{11}

In that way one can make clear that there is no place called Uksermer in Holland (Overmeer: typesetter misread v as r; translator misread r as x - no explanation for the u...) - there was no battle near the island of Akhodia (in the original, Rhodos appeared as Ahodie in stead of Rhodie)^{12} - that a certain Jan van Matenesse was not "a member of the Superior Secret Council" but "president of the Polder Board" (the translator misinterpreted hoogheemraad as hoge geheime raad) - and that the translator did not say Petr Koloma was the king of Spain (a modern editor thought he was)...^{13} Perhaps more practical: one can prevent phantom words entering into historical dictionaries, like Russian gelatence for an exotic kind of textile imported into Holland by the United East Indian Company (the editor of the Russian translation interpreted an unclear scribble as ge in stead of p - platence (or polotence) is a longish piece of fabric, and the Dutch original has lange neus-doecken, which is "long handkerchiefs").^{14} And, hopefully, one can prevent page-long discussions about how the German place name Lehe (in Holstein) became Leher (spelled Leger) in Russian, having found out that the translation Iz Legera vesti pro korolja datckogo ("From Leher news about the Danish king") was wrong, the original having Wt 't Legher van den Coninck in Denemarcken ("From the army of the King in <=of> Denmark") - the translator apparently read: Wt ('t) Legher. Van den Coninck in Denemarcken. (Russians were not used to punctuation and capital letters at that time.)^{15} We also showed that a certain translation from a bilingual brochure was made from the Polish version in it, not from the German one - what may be helpful to future researchers, especially because it apparently is the only remaining translation from a text printed in Polish in fifty years or more (1620-1670...)^{16} of newspaper translations for the tsar.^{17}

Some of those things can be established only if one has found the original. Many conclusions, however, can be reached even without having the source. In any case, I became very much aware of all kinds of errors and mistakes that may happen - in writing, reading, typesetting, copying, translating, editing (then and now),... I learned also that precisely from such errors (and other peculiarities), one often can draw quite accurate conclusions about the language of the original, the words used, etc. That is, one often can find out what it was the translator translated. (Or even, one step further: which idea the original author wanted to express.^{18}) This experience made me try to figure out some possible explanations for the "non-matching" translation pairs in Wim Lucassen's article.

Prima facie - the danger of single translations

Witsen translated one Georgian word (that might have been an expression), giving one expression, often one word, in Dutch. For a presentation of how Georgian words look, that is okay, but we all know that hardly ever one word of a language corresponds to one word in another one. We may suppose that Witsen in his list gave the first Dutch word or expression that came to his mind when seeing the intermediate translation: the prima facie meaning (of the word now lost). Prima facie translation, however, does not work the other way round. English a dean would be Dutch een deken, but Dutch een deken would be English a blanket. The question to be asked is not "What would be the translation of this word?", but "Which word(s) would be translated in this way?"

When someone is asked "what does `a stage' mean?", she or he would probably say something like "a place where actors and musicians do their thing", or, in Dutch, een toneel, een podium. When, however, there is some context, like "at this stage, we cannot decide", the answer probably would be something like "a phase", een stadium, but we cannot exclude an explanation like "(at this) moment", "(at this) time" or "(at) present". The first meaning that comes to mind without context (often the first translation in a dictionary), is the one "on first sight". Nevertheless, different people may give different meanings "on first sight".

If a (e.g. Georgian) word meaning "phase" had been translated as (English) stage, in Dutch it would be rendered by toneel (stage, podium). In that case, when the intermediate translation has been lost, the connection toneel - stage - "phase" would be fairly easy to find, because both meanings of stage (podium, phase) are more or less obvious. Clearly, if the intermediate translation had been time, in Dutch we would see tijd (time). Finding tijd - time - "phase" would be less easy. For present, in Dutch could appear geven (to give). Discovering geven - present - "phase" can be very hard, and even harder when one does not know for certain what was the intermediate language...

Matching non-matching word pairs

Wim Lucassen writes that several not coinciding translation pairs could be matched by supposing French, Russian, Greek or Latin as the language in which the Georgian words of Witsen's list had been translated. An example: Witsen has gras (grass) for tzamali, while tsamali in Georgian means "medicine". In this case at least two languages fit as an intermediate: French herbe and Russian trava allow both interpretations.^{19} They both meet a quite important criterion: one would expect Witsen (or someone else) to translate, out of context, herbe or trava as gras. This enables one to identify tzamali with tsamali. I tried French first (my knowledge of Russian or Latin does not lend itself to free word association). I also tried German, because at the time that language was widely used. For around 1700, English is of course no option. Not knowing Georgian, I could not consider changes in the meaning of the Georgian words, or other meanings not provided in Wim Lucassen's article. [***** Not asked yet! ***** Regarding that, W.L. helpfully made some remarks after this article was written. One will find them in the notes. *****]

Below, I shall give (WD) Witsen's Dutch translation, preceded by its number in the list, with (PF) its (modern) prima facie meaning in English (which may, of course, not be the meaning Witsen had in mind!), then (WG) Witsen's Dutch transcription, (MG) the (modern) Georgian word Lucassen found, in transcription (WG and MG for reference only - this article is not about Georgian!), and (ME) the (modern) prima facie meaning of that word in English.^{20} [***** Not asked yet! ***** I am grateful to Wim Lucassen for checking whether my English rendering of his Dutch translation of the Georgian word is in fact a translation of that Georgian word. That is what this article is all about: the translation of a translation of a word might not be a translation of that word... *****] In most cases I added (IV) a possible intermediate version (which means only: perhaps something like...). (In IV, X = unclear writing, * = non-existent word.)

My impression is that Witsen tried to indicate verbs by "I ..." and (countable) nouns by "a ...", and that this does not mean that "I" or "a" were indeed present in the intermediate translation. It seems also that some expressions have been shortened, e.g. during editing.

It might seem strange, that only a few pairs would have German (or Polish) as an intermediate, but it should be kept in mind that hundreds of translation pairs coincide in such a way that their intermediate language cannot be established. Especially a German intermediate version would often without problems lead to an acceptable Dutch translation. In any case, it is quite certain that Witsen's list was compiled from several lists with different translation languages. Also the frequent mentioning of misspelling, misreading, etc. should not be a surprise: these are the 25 translation pairs out of 900 that still presented a problem!

(As an illustration of the process of finding an intermediate translation, I added some ideas that I had to discard for one reason or another, in special notes indicated by ^{x1} etc.)

Some fairly probable translation pairs

57 WD Ik bevele - PF I command, I order - WG kurutzam - MG grvtsams - ME we believe - IV confier

Even now, bevelen can also mean "to entrust", "to confide" (something to/with someone). (Fr) confier has this meaning of "to confide", but also "to trust", "to have confidence". This is close enough to "to believe". A two-faced verb "to confide" could exist also in other languages.^{x1}

256 WD Een Meir - PF a lake ? - WG kary - MG khari - ME wind (s.)
or             MG kari - ME door - IV baie

Meir in fact was not an old spelling of meer (lake) or meier (bailiff). It was a mooring place or anchorage, an open harbour (physically, not politically), or an inlet (on the coast).
(Fr) baie is "roads" or "(small) bay", and "opening of a door or a window". The combination "roads" (mooring place) and "small bay" more than justifies the translation meir.^{x2}

I did not find a link for the meaning "wind".

267 WD De Hiel - PF the heel - WG pranggi - MG phrangi - ME Frenchman - IV latin

Here, Witsen writes de (the) in stead of een (a). There must be a reason for that. In French there would have been la or le, I presumed.
In the 17th century, in the Levant, the expression Franc, Franque was used for (Western) Europeans in general, or at least for those speaking a Roman language (so-called "Lingua Franca" was an extract of these), or those thought to be Roman Catholics. This might have been the case in Georgia, too. In French, this meaning was expressed by latin (a latin, a (SW-) European). The maker of Witsen's list apparently read la tin. (Fr) (le) tin is "stocks" (in a shipyard) or "keel" (of a ship). In Dutch that is kiel. The typesetter may have misread one letter. (Or there is a misunderstanding: hiel(ing) is a part between the keel and the stern of certain ships, and the lower part of a mast. Witsen would know: he wrote a book on ship-building through the ages with a dictionary of nautical terms...^{21})

303 WD Bequaem - PF able, capable, competent - WG tzinti - MG tsindi - ME pawn, security (object) - IV prêt

(Fr) prêt (old: prest) is "a loan" (which may involve a security), or "an advance (payment)" (which may be a security). As an adjective, prêt is "ready", "able", "in a position" (to do something). Bequaem covers several meanings of prêt. (That prêt is not a security but a loan, does not bother me too much - a description of the agreement (of which both are part) could easily lead to this misunderstanding.)^{x3}

365 WD Een vet Mensch - PF a fat person - WG kuavi - MG gvami - ME a dead body - IV corps (homme)

(Fr) corps is "a dead body", but also "a body" in general. Furthermore it can mean "fatness", "a (fat) belly". The Latin word corpus, when used in Dutch, usually has the connotation "fat/thick", too. (So to say, "a body with body" - what shows, that "body" and "thick" can be linked in other languages, too.).

536 WD Een vuur - PF a fire - WG ekidna - OldG ekhidna - ME (a kind of reptile) - IV couleuvre

Lucassen points out that Old Georgian ekhidna came from Greek. In Greek it meant a snake (a viper). According to Lucassen in Georgian it was the name for "a kind of reptile". (Fr) couleuvre is a non-poisonous snake (e.g. a grass-snake); couleuvre / couleuvrine was a long and thin kind of portable cannon, named after the snake (Dutch, at the time: klover, from which klovenier (soldier of the Civic guard); English, at the time: "culiver"/"culverin"). Dutch een vuur could also be "a gun". I suppose that Witsen in this case meant: "a kind of gun".^{x4}

621 WD Een Wind - PF a wind - WG kvutzi - MG kvithsi - ME foal - IV pouliX

Een wind is in this case not a wind that blows, but an apparatus for winding: "winch" or "pulley". (For "wind", normally not a countable, there would be no article.) (Fr) poulie is "pulley". (Fr) poulain (misspelled: poulin?^{22}) is "foal", pouliche is "filly" (a female foal). The word could have been corrupted in the copying process.^{x5}

655 WD Laken - PF cloth - WG tzochin - MG dzjochi - ME stick (s.) - IV IIIec^ = shest"

Russian sherst' (wool, woollen fabric) and shest" (rod, pole) would in the handwriting of the time look very similar.^{23},^{x6}

657 WD Ik snork - PF I snore - WG kantzanrep - MG gavdzjavreb - ME I imitate - IV affecter

Ik snork could (and still can) also mean "I boast", "I brag", "I pretend". The meaning "I pretend (to be what or who I'm not)" is close enough to "I imitate". The French verb affecter (to boast, to pretend, to do as if,...) might be a good candidate, but of course some other language could have been used.

667 WD Ik Speele - PF I play - WG tvare - MG thvare - ME moon - IV crossant

(Fr) croissant is the "crescent" (the moon). (Fr) crossant is from crosser that means "to play/strike the ball (or puck,...)" (in Dutch: "de bal spelen") - if only a crosse (a curved stick) is involved, like in golf, hockey,... Indeed at the time a similar sport (Dutch: "kolven", (Fr) "crosse") was quite popular. As a translation, I would expect "Ik speele den bal", but "den bal" might have been lost during editing.

672 WD Een Rug - PF a back (of the body) - WG ate - MG ade - ME get up!, stand up! - IV rücken

Here I did not find any correspondence through French. (Closest would be: do(r)s is "back", dehors! is "get out!") But German Rücken is "back (of the body)", and rücken as a verb can mean "move up", "move over", "go and sit/stand somewhere else". Close enough.

694 WD Een Sweep - PF a whip - WG piri - MG piri - ME mouth, face - IV Zug

German Zug is "stroke (line)", "feature of the face", and, among other things, horses pulling a cart. It is not a whip. Sweep could be a misprint for Streep (line, dash) (with the writing of the time, that could be possible), and that would be a prima facie translation of Zug. However, according to Witsen's nautical dictionary^{21}, sweep is a rope for drawing or fixing a cable. That, too, in German is a Zug (or Zug-Leine). The relation between "line in the face" and "line", and between "line" and "to draw" could exist in various languages: also French trait (old: traict) is a feature of the face, a stroke, and a rope for pulling (if not for pulling a cable).

802 WD Keyen aen den Oever - PF Stones on the shore/bank - WG mekvape - MG mekhvabe - ME copper-smith - IV gravXXr

(Fr) gravier is "gravel", "pebbled sand" ("pebbles on the beach..."). (Fr) graveur is an "engraver". Apparently the Georgian copper-smith embellished his products by engraving them. And someone misspelled, misheard or misread. (There are professions with -eur and with -(i)er...) Given gravier, then keyen aen den oever might be the first thing that comes to someone's mind.

815 WD Hoozen - PF (an out of date kind of) trousers - WG chelsachos - MG chelsachothsi - ME towel - IV braie

(Fr) braie (old: braye) was "swaddling clothes", now (GB) "napkin", (US) "diaper". This may well correspond to "towel". Also, braie was (an old Gallic form of) "trousers". Dutch hoozen was "trousers", too. (German Hose still is.)^{x7}

832 WD Een uitgieting - PF a pouring out - WG tzkarat - MG thsjkharad - ME fast (adv.) - IV coulé

Any pouring can be fast, in various languages.
(Fr) couler is "to pour/flow (out)", or "to cast (metal)"; (comme) coulé is "poured out", "thrown", "cast", which can mean "fast". Witsen has een for a countable noun. He may have understood (un) coulé "a cast(ing)".

Pairs with question marks, but still possible

402 WD Een Zyde - PF a side - WG metzepari - MG medzebari - ME hunting dog - IV Seidenhund

German Seidenhund (`silk-dog') is e.g. "a spaniel", which is a hunting dog. In Dutch that is een zyde-hond. However, I have not yet found the short form een zyde. Did the dog get lost during make-up?^{x8}

744 WD Ik schiet met Pylen - PF I shoot with arrows - WG taperte - MG daberdi - ME you have become old - IV *acher

(Fr) archer (to shoot with arrows) and agé (aged, old) may sound quite similar, when spoken by non-French speakers. Also, in writing g and ch could look similar.^{24} If acher were written or read, the interpretation archer is quite possible.

758 WD Aengenaem - PF pleasant, agreeable - WG achali - MG achali - ME new - IV *chermant

(Fr) germant is "germinating", "beginning to develop", (i.e.: "new"); charmant is "charming", "(very) pleasant"; cherement is "dearly" (both "cordially" and "costly"). All three, when spoken by non-French speakers, may sound very similar. Also, in writing g and ch could look similar.^{24} If chermant were written or read, the translation aengenaem, that covers both cherement and charmant, would be a good solution.^{x9}

790 WD Een Blad - PF a leaf - WG mokepini - MG mokepili - ME chopped (into pieces) - IV hache ?

(Fr) haché is "chopped in pieces" (wood, meat,...); hache is an "axe" (or some other chopping device). However, blad is the "blade" of the axe, not the axe itself. It was used pars pro toto as a general word for a chopping/cutting instrument.^{x10}

822 WD Verhooging - PF elevation, augmentation - WG vitamasep - MG vithamasjeb - ME I shall play - IV hausse (d'enjeu)

If in Georgian the game is about money, and "I'll play" means: "(I'll not pass, but) I'll raise my stake", in various languages the intermediate translation could result in verhooging (raising). (Fr) hausse d'enjeu (raising of the stake) could have been possible.^{x11}

The remaining pairs (without a real solution yet)

439 WD Goud - PF gold - WG sugli - MG sjughli - ME - quarrel - IV zlot ??? / Groll ???

Polish zloto (gold), zlot (flying at each other(?)).
German Groll (grudge, anger) could under circumstances be read as Golt (for Gold (gold)).^{25}

496 WD Een Bekende - PF an acquaintance - WG kamchmari - MG gamchmari - ME dry - IV sXchant ? / *saichant ?

Een bekende was not only "someone who is known", as it usually is today, but also "someone who knows" (who knows about things, is familiar with things). In French, sachant is "knowing", and un sachant is "a knowing / knowledgeable man", "an expert" (e.g. in court) (in contrast with un savant, who is "a learnéd man", "a scholar"). (Fr) sec/sèche is "dry"; séchant is "drying" (both making dry and becoming dry). If the second letter could not be read, the misunderstanding could happen. Also if saichant were written: sounding like sé-, looking like sa- (a and ai alternate in the verb savoir). (But why "drying" if the word is "dry"?)

509 WD Laken - PF cloth - WG sultzota - MG sul thsota - ME a little bit - IV petit pXX ???

(Fr) un petit peu is "a little bit". (Fr) drap à petit poil would be "short piled cloth". I did not find the shortening petit-poil, but perhaps Witsen heard it among cloth merchants. The word peu should then have been read as poil (something in between would not do if no cloth were mentioned) - what under circumstances could have happened.^{26}

827 WD Hooy - PF hay - WG metzirepa - MG metsireba - ME will be sacrificed to me - IV soin ???

(Fr) soin is "care", "dedication". Most probably the word was written with a long s (like an f without horizontal stroke). It could have been misread as foin, i.e. "hay". To what degree "dedication" corresponds to "sacrifice"?^{x12}

844.1 WD Een Vlo - PF a flea - WG magali - MG maghali - ME high
or            MG maqali - ME chafing-dish, brazier - IV poXl ???

(Fr) puce (old: pulce) is "flea". (Fr) pou is "louse"; pouilleux is "full of lice". The old spelling for "louse" and "lice" is poulx. Latin pulex, however, is "flea". (Fr) poêle (old: poële, poelle, poile) is "frying pan" (then also: on a tripod over the fire) or "stove".^{27} A frying pan over burning coal or wood reminds of a brazier. A corrupted form like pouil(e) could be misinterpreted as luis (louse), or even (because of the Latin form) as vlo (flea).

For the competing meaning "high" I have not found a clue (do fleas jump high?) [but see note!] ^{28}

Caution

The last remark (fleas jumping high) shows that this procedure of back-translation should be applied with caution. The translations on both sides of the language barrier must be reasonable. The intermediate expression does not have to be a prima facie translation or even a true translation of the Georgian word (context may have lead to a different interpretation, and the translator may have paraphrased the notion), but the expression Witsen gives, should be a possible prima facie translation of the intermediate version (for him, I fear, there often was no context at all). Only then we may conclude that Witsen's Georgian word probably is identical with (or closely related to) the one Lucassen found.

Hocus-pocus

This way of trying to establish the identity of words may sometimes look like hocus-pocus - one may think it is so fuzzy that it hardly can be accepted as proof. Indeed, there is no proof. But one should only try to take a pair of unrelated words that sound (nearly) the same in one's own language, and then find a word in a pre-defined (not intimately related) language, which could serve as a translation of one of the homophones, and which would be translated as the other one. In my experience, such homophones or near-homophones are very rare. In this case, if the pair is really a pair, the assumption is that the Georgian words meant by Witsen and found by Lucassen, are the same or closely related, and we know for certain that an intermediate translation existed. This makes me confident that finding a possible intermediate version is helpful in deciding that the two word forms represent the same word.

For more than half of the 25 "missing links" I am sure that the intermediate translation I imagined, will remove any doubt about the identity of Witsen's word. Several other explanations that depend on the unverifiable writing form of certain letters, to me - and hopefully to others - are so convincing that I am confident to say that now 99% of the words in Witsen's Georgian list have been identified. A nice number. I am pleased that I could contribute one, or even two, percent.

Notes

^{1} For an impression on Nicolaas Witsen, see Marion H. Peters' page on the Internet: http://www.xs4all.nl/~mhpeters/.
^{2} "Tartary" was everything from, roughly, the river Volga to Japan, from the Himalaya to the Northern Ice Sea. It has little to do with nowadays Tatars or Tatarstan. Witsen's work is at least famous among the peoples he has written about, because often he was the first one, or one of the first, to do so.
^{3} Lucassen 2003a.
^{4} Lucassen 2003. On p. 169: "Finally there now remain 25 word pairs with an ununderstood correspondence of meaning. [...] Who will help puzzling?"
^{5} Such information can be useful, e.g. because at the time not only Russian, but also French, German and Dutch would (or at least could) be written each in a different type of script. (Around 1700, the `modern' Italian script was widely used in France, but had not yet fully conquered the Netherlands and Germany. The old Russian script, very different from the one used today, was in its last days, but tsar Peter's script reform was yet to come.) This situation quite often lead to script mistakes, like using a letter form belonging to script A while writing a word in script B, or writing a quoted word of another language in the `wrong' script. (If I quote a Russian word Puma - how the reader should know whether I mean, in English sounds, p-oo-m-ah or r-ee-t-ah? And if I do not say that it is Russian - who would, seeing Puma, think of Margy?) The problems poulin / poulie (see note ^{22}) and peu / poil (see note ^{26}) might be examples of that. The use of another language than one's own of course may lead also to other mistakes. [In the text of this article, with the exception of IIIec^ in word number 655 and in note ^{23}, Russian is always transcribed in Latin script!]
^{6} I'm not aiming at, nor capable of, repeating Lucassen's project... I hope this article is in time for him to make use of my findings. 879 is rhetoric - the true number is 875: see note ^{19}.
^{7} Her name is Ingrid Maier. See the references below.
^{8} e.g. Maier 1997. Others who asked those kind of questions: e.g. Waugh 1979, Schibli 1988, Dem'janov 2001.
^{9} See e.g. Maier 2003.
^{10} cf. Maier 2003a, or the introduction to Maier 1997.
^{11} Some examples of holes `filled' in that way, will be shown in Maier 2005?. In a few cases, a copy of the newspaper without a hole in the news item concerned, was found later in an other archive. For smaller holes (a few words) the back-translation showed mostly to be accurate or near-accurate; for larger holes it was in two cases quite acceptable and in one case rather far from the real contents.
<sup>{12}</sup> Maier 2003a, 56.
<sup>{13}</sup> For Uksermer, hoogheemraad, Koloma, see Maier & Pilger 2001: 227, 227, 226.
<sup>{14}</sup> Maier & Pilger 2003, 207.
<sup>{15}</sup> Maier & Pilger 2003, 211; Dem'janov 2001, 43.
<sup>{16}</sup> We have not yet studied the newspaper translations after 1670.
<sup>{17}</sup> Maier & Pilger 2003a.
<sup>{18}</sup> cf. Pilger 1993 (about translating poetry).
<sup>{19}</sup> Lucassen 2003, 171, where he gives a few other examples. For the pairs he gives as mon temps > montagne ("time" > "mountain") and argenter > argenté ("it costs" > "silvered tin") I think of montre > mont and (il) nous faut (de l') argent > nouveau[!] argent. See note <sup>{6}</sup> for why I do not elaborate on that.
<sup>{20}</sup> WD, WG, MG as in (and ME after) Lucassen 2003, 169-170. MG in Georgian script one will find there. For a better comparison with Witsen's spelling, I left Lucassen's transcription in its Dutch style. (Consider, if you like, j as an accent on the previous letter.)
<sup>{21}</sup> Witsen's nautical dictionary: Witsen 1671, pp. 481 ff.
<sup>{22}</sup> In some styles of Old German handwriting, e and n looked similar. If written in such a script, poulin could look like poulie. But writing a French word with German letters was `not normal'. (See note ^{5}.)
^{23} Russian: sherst' (wool, woollen fabric), shest" (rod, pole). In Russian 17th-century chancellery handwriting (skoropis') the r and the t would have been written with signs above the other letters (vynosnye bukvy), and the raised t would include the soft sign ' or the hard sign " or any other letter, or nothing. One word would look like IIIec^ (shest"), the other like IIIèc^ (sherst'). There may (not necessarily) have been an unintended dash over the e in shest", making it look like sherst'. If the last letters had been written as normal letters on the line, the signs for ' and " could look the same. (Cf. Beljaev 1911). (The style of handwriting in Witsen's source is not known. For Russian, at the time it would have included raised r's.)
^{24} In (Dutch or French) handwriting a g could look like the combination of a c plus a long h. (Cf. Horsman 1986, 15).
^{25} Polish: zloto (with a dash through or over the l, perhaps neglected) is "gold"; zlot (without dash) is "birds flocking together", "a birds' assembly" (many birds getting down in one place, before flying off together), "flocking, running, mobbing together" (many people getting together in one place (fast); modern: "a flash mob"), or simply "a gathering" (now e.g. of scouts). Could it be a less friendly meeting ("to fly at one another")? Is this the kind of quarrel meant? Could there have been Polish translations? Witsen had good Polish contacts, but did they contribute to the Georgian list?
In Old German handwriting, capital letters were often elaborate, but not always very clear. It is possible that the German word Groll (grudge, anger) was written in such a way, that it could be read as Golt (gold): what was written as an r could be read as a part of the capital G (cf. Horsman 1986, Nr 88) - that the l would be read as a t is less surprising when a word Goll does not exist. (The same things may happen in modern handwriting.) There is not a great difference between "to have a quarrel with someone" and "to be angry at someone" or "to bear a grudge against someone".
^{26} In Old Dutch handwriting e could indeed look like o and u (normally written as ú) like il, but, generally, around 1700, French words, even in a Dutch context, would not be written in Old Dutch script. (That might precisely be the reason why it was misread...) Chances in German script are less. (See note ^{5}.) The kind of handwriting in Witsen's source is unknown.
^{27} In fact la poêle is a frying pan, and le poêle is a stove. Both are from Latin patella. In Spanish, paella - a meal prepared in a frying pan over burning coal or wood - was named after the (Catalonian) frying pan.
^{28} In sloppy handwriting poe(l)le (frying pan, stove) could be read as pu(l)ce (flea). In equally sloppy Russian skoropis' handwriting, vysoko (high) could be read as bloxa (flea)! - i.e.: no solution!

Discarded ideas

[It is not surprising that sometimes there are more candidates for an intermediate translation if those have parts of their semantic fields in common, but sooner or later a reason appears why the one fits and the other does not.]

^{x1} [discarded:] Before, (Fr) commander (and Dutch bevelen) had also the meaning "to commend" (e.g. one's soul to the Lord); quite close to "to believe". There should have been a hint, e.g.: commander (son âme).
^{x2} [discarded:] (Fr) ouverture (opening) also meant an opening to sea - as well as the opening of a door. However, ouverture would be translated as meir (in stead of opening) only if there were extra information, like ouverture (porte) (door) written or read as (port) (harbour). (Spelling was not standardised at the time!)
[discarded:] A quite simple explanation is porte ouverte (open door) written or read as port ouvert (open harbour). In that case, however, one would wonder why Witsen did not simply write open haven.
^{x3} [discarded (but not entirely):] (Fr) habile, like its obvious Dutch translation bequaem is "able, competent" and "authorized, entitled". I see a possible connection between "entitled" and "a security". (Titles (rights) can be given as security for a loan: prêt sur titres.)
[discarded (but not entirely):] At the time, bequaem could also mean "comfortable", "at ease" (like now in German), so any intermediate translation meaning "secure" (adj.) would fit. However, the Georgian word indicates an object.
^{x4} [discarded:] A salamander is traditionally associated with fire. It is not a reptile but it looks like a lizard. Could the intermediate translation have been "a fire-lizard" or something like that? (Language?) (In German there is e.g. Feuermolch (a water salamander) and Feuernatter (a snake, i.e. a reptile).) The "lizard" (or "snake") part should then have been lost on the way. (Een vuur could not be "fire" as a phenomenon, only "a fire".)
[discarded:] "Fire" meaning "gunfire" ((Fr) tir or tiraillement) would not be likely: Witsen probably would not have used the article een in that case. A name of a reptile sounding like tir I have not found.
[discarded:] [It may be that Witsen still distinguished vier (fire with flames) and vuur (gun, gunfire).]
^{x5} [discarded:] (Fr) poulain is "a foal", but also a plank (or two beams) for rolling (wine) barrels up or down - there might be some kind of winch involved. I doubt it would be translated as een wind.
^{x6} [discarded:] (Fr) ramie (Dutch: ramie or ramee) is a textile fibre from a Chinese or East Indian nettle; ramille (pronounced about the same) is "a small twig"; ramée is "leaved branches" (cut off or overhead). Neither a twig, nor a branch with leaves is suitable to be "a stick", and any kind of (nettle) muslin (Dutch: neteldoek) is too thin to be called laken. There may have been obscurity about the French words for sticks, branches and twigs - there would be no misunderstanding about muslin or laken in 17th-century Holland in the surroundings of Nicolaas Witsen, I think...
[discarded:] Dutch laken can and could also be a (bed-) sheet (that could be thinner and not woollen). But then Witsen would have written een laken.
[discarded:] (Fr) pique is "a pike", i.e. a pointed stick; piqué is a specially woven cotton fabric. Dutch laken in principle is a felted woollen fabric.
^{x7} [discarded:] (Fr) chausse is, now, a longish piece of textile, which (French) professors wear over the left shoulder of their robe or toga; its colour shows the faculty they belong to. Apparently they used to wear their stockings over their shoulder, for the word can and could mean only "stocking" or a funnel-formed felt "filter-bag" (for (spiced) wine, etc.). The plural form, chausses, is a kind of (tight) trousers of the time, in Dutch: hoozen. This makes sense only, if the Georgian towel is or was a tube-formed piece, or a piece of fabric like before was being used for winding around the legs or the feet (a kind of "puttee"), or if its fabric is or was of a type used for filtering.
^{x8} [discarded:] I first thought of zij(de)-hond (side-hound), like existing zij-geweer (zy(d)-gheweer) (side-arms: personal weapon(s), often a sabre), in German that might be Seitenhund like existing Seitengewehr. But then I found Dutch zijdehond and German Seidenhund for "silk-dog" (e.g. a spaniel) - Dutch zijde (zyde) is both "side" and "silk".
[discarded:] Een zyde could also be "a side (of bacon)". (Fr) flèche has that meaning, alongside with "arrow" and "lightning". Could it be "a dog fast as an arrow" or "fast as lightning"? However, flèche would be translated as pijl (arrow) if there was no extra hint.
^{x9} [discarded:] French frais and German frisch, as English "fresh", can mean both "pleasant" (e.g. weather in summer) and "new". (In Dutch the word differentiated: "fris" (for the weather) and "vers" (for the milk) - a case of metathesis like Wim Lucassen encountered many, comparing Witsen's Georgian words with modern forms). (Lucassen 2003, 170ff.) However, aengenaem in my eyes is not an obvious translation for frais or frisch.
^{x10} [discarded:] (Fr) feuillet is "a small leaf", but could also be "a thin piece of", "a slice". This could be a piece chopped off (but not, normally, in French). If feuillet were written, one would expect blaadje or bladje (small leaf), but in fact feuillet is also a leaf (two pages) of a book, in Dutch: blad.
[discarded:] (Fr) tranché can be "cut up" (normally with a knife), while tranche is a flat piece cut off or sawn off. It could be a slab of marble, or a wooden board. Then the translation blad would be right. But that is not a prima facie translation for tranche.
^{x11} [discarded:] If the Georgian expression relates to theatre or music: (Fr) estrade (raised platform, podium) would be translated as verhooging, because the word is not used for a stage in the theatre. But it certainly is the right word for a stage outside the theatre, where one would perform a play, music, etc. However, Witsen did not use the article een, so, if he was consistent in that, he probably did not have an object in mind.
[discarded:] [An other word for "augmentation" in Dutch is opvoering. That can also mean "playing a play". If any intermediate translation had been rendered as opvoering, and if this during editing became verhooging, this would give another explanation. But only if.]
^{x12} [discarded:] (Fr) faucher is "to mow"; fauchage is "(the act of) mowing", but also "the result of mowing", i.e. "hay". (Fr) fauche is also the mown grass lying on the field for drying (and becoming hay). (Fr) faucher can also be "to destroy, to kill (many or all)" (about Death, illnesses,...), or "to shoot in various directions" (for large scale killing). Both might relate to "to sacrifice". In other languages, too, "to mow" could relate to "to kill (many)". But is this the kind of sacrifice meant?
[discarded:] (Fr) meule is "haystack" (and would probably have been translated as such), immoler is "to sacrifice", but I do not see a connection. (The verb mollen in Dutch (received slang) means "to kill", "to ruin", mollement (a quasi-French noun in Dutch, unrelated to the French adverb which means "tenderly") is documented for "killing by stabbing" in 1731.) (Cf. Endt 1974, 90, where it is linked to Romany muló (dead).)

Dictionaries

The various high-quality dictionaries made by the editorial staff of (former) Kramers' Woordenboeken, especially the once-only paperback Editie 2000 published by HMP, Amsterdam under the titles Het complete woordenboek... (N-F & F-N, N-D & D-N, N-E & E-N, Nederlands), have been a great help. (Hardcover editions with apparently the same contents were later published by Standaard, Antwerpen under the titles Standaard handwoordenboek ...). The 1952 edition of Nouveau petit Larousse illustré contains a surprising number of old French words - for Dutch the same holds for Verschueren Groot encyclopedisch woordenboek (edition 1996) and its predecessor Verschuerens modern woordenboek (edition 1965).

17th-century meanings have been checked in (i.a.):

Dutch
WNT = Woordenboek der Nederlandsche taal op cd-rom, Leiden 2000. (paper: 1864-1998)
[Covers the Dutch language from 1500 to 1920. It includes information from 17th-century (and older) dictionaries, which is hard to locate in the yards stretching paper edition of `the world's largest dictionary'.]

French
DAF = Dictionnaire de L'Académie Française, Premiere édition (1694)
Jean-François Féraud: Dictionaire critique de la langue française (1787-1788)
both in: Dictionnaires d'autrefois
Internet: http://www.lib.uchicago.edu/efts/ARTFL/projects/dicos/onelook.html
and - because Witsen's contributors probably worked earlier and did not follow the preferences nor the spelling of the Académie - using various old spellings:
Dictionnaires de la Renaissance (Jean Nicot: Thresor de la langue françoyse (1606), and earlier dictionaries).
Internet: http://www.chass.utoronto.ca/~wulfric/dico_tactweb/tiden.htm

German
Googling the few words concerned gave enough information.

References

Beljaev, I.S. (1911). Prakticheskij kurs izuchenija drevnerusskoj skoropisi. Moskva. [A practical course for learning the old Russian skoropis' script.]
Dem'janov, V.G. (2001). Inojazychnaja leksika v istorii russkogo jazyka XI-XVII vekov. Problemy morfologicheskoj adaptacii. Moskva. [Lexical units from other languages in the history of 11th- to 17th-century Russian. Problems of morphologic adaptation.]
Endt, Enno [1974]. Bargoens woordenboek. Kleine woordenschat van de volkstaal. 2e druk. Amsterdam. [Flash dictionary. Small vocabulary of the vulgar tongue.]
Horsman, P.J., e.a. (1986). Schriftspiegel. Nederlandse paleografische teksten van de 13e tot de 18e eeuw. Zutphen. [Script-mirror. Palaeographic texts from the Netherlands, from the 13th to the 18th century.]
Lucassen, Wim J.M. (2003). `Omzetten en omzetten is twee', in: Die het kleine eert, is het grote weerd. Festschrift voor Adrie Barentsen (= Pegasus Oost-Europese Studies 1). Amsterdam. pp. 167-175. [Translation and translation (transposition) are two.]
Lucassen, Wim J.M. (2003a, to appear). Witsen's Dutch-Georgian word list in Noord en Oost Tartarye (Amsterdam 1705, pp. 506-515) provided with Modern Georgian, Dutch, English and Russian equivalents.
Maier, Ingrid (1997). Verbalrektion in den "Vesti-Kuranty" (1600-1660). Eine historisch-philologische Untersuchung zur mittelrussischen Syntax (= Acta Universitatis Upsaliensis. Studia Slavica Upsaliensia 38). Uppsala. [The use of cases with verbs in the "Vesti-Kuranty". A historical-philological research on Middle-Russian syntax.]
Maier, Ingrid (2003). `Amsterdamer und Haarlemer Zeitungen ("Couranten") des 17. Jahrhunderts im Niedersächsischen Staatsarchiv zu Oldenburg', in: Gutenberg-Jahrbuch 2003. Mainz. pp. 170-191. [Amsterdam and Haarlem newspapers ("Couranten") of the 17th century in the Lower-Saxonian State Archives at Oldenburg.]
Maier, Ingrid (2003a). `Die Übersetzungen westeuropäischer Zeitungen am Moskauer Gesandtschaftsamt von 1660 bis 1670: Zur ersten Ausgabe der Vesti-Kuranty mit Paralleltexten', in: Swedish Contributions to the Thirteenth International Congress of Slavists, Ljubljana, 15-21 August 2003 (= Slavica Lundensia Supplementa 2). Lund. pp. 51-74. [The translations of West-European newspapers at the Moscow Diplomatic Chancellery from 1660 to 1670: About the first edition of the Vesti-Kuranty with parallel texts.]
Maier, Ingrid (2005?, in preparation) = V-K VI/2 = Vesti-Kuranty, 1660-1670 gg., Chast' II, Inostrannye istochniki k russkim perevodam. Moskva. [Vesti-Kuranty 1660-1670, Part 2, Foreign sources for the Russian translations.]
Maier, Ingrid & Pilger, Wouter (2001). `Second-hand translation for tsar Aleksej Mixajlovich - a glimpse into the "Newspaper workshop" at Posol'skij prikaz (1648)', in: Russian Linguistics 25. pp. 209-242.
Maier, Ingrid & Pilger, Wouter (2003). `VOC-ladinglijst vertaald voor de Russische tsaar (1667)', in: Die het kleine eert, is het grote weerd. Festschrift voor Adrie Barentsen (= Pegasus Oost-Europese Studies 1). Amsterdam. pp. 191-213. [United East Indian Company cargo manifest translated for the Russian tsar (1667).]
Maier, Ingrid & Pilger, Wouter (2003a). `Polnische Fabelzeitung über Sabbatai Zwi übersetzt für den russischen Zaren (1666)', in: Zeitschrift für slavische Philologie 62:1. [Polish news fiction on Sabbatai Zwi translated for the Russian tsar (1666).]
Pilger, Wouter F. (1993). `Chiu originalo estas traduko - chiu traduko estu re-traduko', in: Survoje 16 (Sevastopol'). pp. 10-28. [Every original is a translation - every translation should be a renewed translation.]
Schibli, Roland (1988). Die ältesten russischen Zeitungsübersetzungen (Vesti-Kuranty), 1600-1650. Quellenkunde, Lehnwortschatz und Toponomastik. (= Slavica Helvetica 29). Bern. [The oldest Russian newspaper translations (Vesti-Kuranty), 1600-1650. Sources, loan-words and place names.]
Waugh, D.C. (1979). `News of the False Messiah: Reports on Shabbetai Zevi in Ukraine and Muscovy', in: Jewish Social Studies 41, 3-4. pp. 301-322.
Witsen, Nicolaes (1671). Aeloude en hedendaegsche scheeps-bouw en bestier: ... Verrijckt met een reex verklaerde Zee-mans spreeck-woorden en benamingen. ... Amsterdam. (Reprints: Alphen a.d. Rijn, 1979; Franeker/Alphen a.d. Rijn, 1994.) [Ancient and modern ship-building and operation: ... Enriched with a series of seamen's proverbs and terms, with explanation. ...] The nautical dictionary is on the Internet: http://pc-78-120.udac.se:8001/WWW/Nautica/Etymology/Dutch/Witsen(1671)_p481.html