Some time ago there was a discussion about the ordinal and cardinal numerals in this forum. The focus of discussion was on the numerals. However, the function word te7 第 was not discussed. It is interesting to note that the usage of te7 in the construct of ordinals in Taiwanese/Hokkien is similar to the the in English and der in German:
Taiwanese/Hokkien:English:German
te7 it 第一:the first:der (die, das) erste
te7 ji7 第二:the second:der (die, das) zweite
te7 sann 第三:the third:der (die, das) dritte
etc.
It is especially noteworthy that Taiwanese/Hokkien lacks the use of the definite article, and yet the te7 第 as a function word works like the 'the' in English and 'der' in German, and sounds similar too!
Heruler
The ordinals in Taiwanese, English and German
Hi Heruler,
I'm sorry, but this last example really doesn't tie. The 'the'/ 'der' in German is NOT the element that transforms a cardinal 'four' into an ordinal 'fourth' -that would be the '-th' in English and the '-te' in German. 'the' is just the definite article and it's clearly not used the same way in the two languages - 'te' and 'di4' are never used as definite articles in Chinese. If you wanted to use this example, you'd have to compare the '-th' ending with the Chinese 'te'/ 'di4'.
Before I'd look at the formative elements like -th, I would start with the numbers themselves, though. For English and German (part of the family of Germanic languages, itself a subfamily of the Indo-European language family) this would be (I'll add some more examples from the IE languages + Mandarin/ Hokkien for comparison):
Eng. Ger. Lat. Rus. Greek Man Min
One Eins Unum Odin Ena Yi Chit/ It
Two Zwei Duo Dua Duo Er Zi/ Nng
Three Drei Tri Tri Tria San Sa"
Four Vier Quattuor Tshetyre Tettares Si Si
You'll probably agree that the IE numbers have much more similarity amongst themselves than with the Chinese numbers.
But similarity itself is not enough. To really establish a proven connection, you need to show predictable sound correspondences. You'll notice, for example, that an English 'two' seems to correspond to a German 'Zwei' or 'zwo', depending on context. Let's look at some more examples:
Engl.: toll German: Zoll
Engl.: ten German: Zehn
Engl.: tap German: Zapfen
Looks like there is a rule here: English /t/ at the beginning of words is German /z/. Let's see - if this is a good rule, I should be able to use it to make predictions: German 'zu' is ... yes, English 'to'. German 'Zinn' is ... yip, English 'tin' and so forth. Looks like a good rule.
And then you'll find some words that do not seem to fit the rule. 'Table' for example (German 'Tafel'). And Telephone (German 'Telephon'). And 'technique' (German 'Technik'). If you look hard enough, you'll find a second rule, namely that all these words that do not follow the 't' <-> 'z' rule have been introduced after the 't' -> 'ts = z' shift took place and come from Latin or Greek - they are loanwords of a later date. In other words, you can refine your correspondence rule to exclude all words of direct Latin or Greek origin. That's still easy enough. If you make a rule for one word it's probably a waste of time - but if you can show sound correspondences for dozens of words then you're probably looking at something real.
You can do the same within the Chinese family. Practically all words in the Mandarin second tone will be Hokkien 5th tone. That's because they both come from the ancient lower ping2 tone. Most words with a j- at the beginning in Mandarin will have a k- at the beginning in Hokkien, because they originally had a k- sound in Middle Chinese. And so on.
Such are useful correspondences that prove a relationship between languages. If I can establish sound correspondences that apply for a large number of words. If you can do this for Chinese and Germanic that would be great. It would have to be useful, verifiable rules though, not random sound correspondences (you can find a handful of words that are the same in any two languages on this planet, especially if you stretch the meanings - that's just statistics). Start with the numbers, the personal pronouns and some common objects etc. - if you can truly find reasonable correspondences you might be in for the next Nobel prize.
BTW: Why would you compare Germanic and Chinese - shouldn't it be Indo-European and Chinese? Or are you suggesting that Chinese branched off the Germanic part of the Indo-European family tree after the Germanic languages became a branch of their own (about 5,000 years ago)?
Best regards,
Aurelio
I'm sorry, but this last example really doesn't tie. The 'the'/ 'der' in German is NOT the element that transforms a cardinal 'four' into an ordinal 'fourth' -that would be the '-th' in English and the '-te' in German. 'the' is just the definite article and it's clearly not used the same way in the two languages - 'te' and 'di4' are never used as definite articles in Chinese. If you wanted to use this example, you'd have to compare the '-th' ending with the Chinese 'te'/ 'di4'.
Before I'd look at the formative elements like -th, I would start with the numbers themselves, though. For English and German (part of the family of Germanic languages, itself a subfamily of the Indo-European language family) this would be (I'll add some more examples from the IE languages + Mandarin/ Hokkien for comparison):
Eng. Ger. Lat. Rus. Greek Man Min
One Eins Unum Odin Ena Yi Chit/ It
Two Zwei Duo Dua Duo Er Zi/ Nng
Three Drei Tri Tri Tria San Sa"
Four Vier Quattuor Tshetyre Tettares Si Si
You'll probably agree that the IE numbers have much more similarity amongst themselves than with the Chinese numbers.
But similarity itself is not enough. To really establish a proven connection, you need to show predictable sound correspondences. You'll notice, for example, that an English 'two' seems to correspond to a German 'Zwei' or 'zwo', depending on context. Let's look at some more examples:
Engl.: toll German: Zoll
Engl.: ten German: Zehn
Engl.: tap German: Zapfen
Looks like there is a rule here: English /t/ at the beginning of words is German /z/. Let's see - if this is a good rule, I should be able to use it to make predictions: German 'zu' is ... yes, English 'to'. German 'Zinn' is ... yip, English 'tin' and so forth. Looks like a good rule.
And then you'll find some words that do not seem to fit the rule. 'Table' for example (German 'Tafel'). And Telephone (German 'Telephon'). And 'technique' (German 'Technik'). If you look hard enough, you'll find a second rule, namely that all these words that do not follow the 't' <-> 'z' rule have been introduced after the 't' -> 'ts = z' shift took place and come from Latin or Greek - they are loanwords of a later date. In other words, you can refine your correspondence rule to exclude all words of direct Latin or Greek origin. That's still easy enough. If you make a rule for one word it's probably a waste of time - but if you can show sound correspondences for dozens of words then you're probably looking at something real.
You can do the same within the Chinese family. Practically all words in the Mandarin second tone will be Hokkien 5th tone. That's because they both come from the ancient lower ping2 tone. Most words with a j- at the beginning in Mandarin will have a k- at the beginning in Hokkien, because they originally had a k- sound in Middle Chinese. And so on.
Such are useful correspondences that prove a relationship between languages. If I can establish sound correspondences that apply for a large number of words. If you can do this for Chinese and Germanic that would be great. It would have to be useful, verifiable rules though, not random sound correspondences (you can find a handful of words that are the same in any two languages on this planet, especially if you stretch the meanings - that's just statistics). Start with the numbers, the personal pronouns and some common objects etc. - if you can truly find reasonable correspondences you might be in for the next Nobel prize.
BTW: Why would you compare Germanic and Chinese - shouldn't it be Indo-European and Chinese? Or are you suggesting that Chinese branched off the Germanic part of the Indo-European family tree after the Germanic languages became a branch of their own (about 5,000 years ago)?
Best regards,
Aurelio
Dear Aurelio:
Very happy to see that you have responded to my posting on the Taiwanese (Tw) ordinals. Your criticism is well founded in linguistics, and I felt your commentary has taken the forum to a new height.
Your penetrating questions truly touched the heart of the issue. I will try to respond to them one by one. Let me address the first question first. What you said is right about the ordinals in Germanic languages whose characteristic element is missing in Taiwanese. Let me explain the missing link in a slightly expanded fashion. The ordinals in Germanic languages are composed of a definite article (def.art.) and a numeral suffixed with a morpheme for the ordinal. To form the ordinals other than the first, second, and third, Indo-European (IE) has an affix * + t. This was retained in Germanic (Gmc.) in cases like *sexst+ ‘6th’, but generally by Verner’s law had changed to Gmc. * +θ or * +ð. (Reference: Joseph B. Voyles, 1992. Early Germanic Grammar. Pre-, Proto-, and Post-Germanic Languages, Academic Press, San Diego, CA. p. 248.) In Old Norse (ON), the affix is either -di or -ði. Only in a few cases is the affix -ti, e.g., sétti ‘6th’.
When I compared Tw words with Germanic in general, I found the following interesting relationships. (1) If the Gmc. words are monosyllabic and the words end in a consonant cluster, the consonants are either dropped or simplified in Tw words. (2) If the Gmc. words have more than 1 syllable, the first syllable is usually retained, and the rest is “lost”. Only occasionally do we find 2 syllables are retained, e.g., ON teina (basket made of twigs) : Tw tîn-nâ 藤藍 ; ON amma (grandmother) : Tw a-má ; ON orð-lof (praise) : Tw o-ló (praise).
In the next posting (in a new thread), I will present a pattern of sound correspondence between ON and Tw to illustrate Relationship (1). Hopefully I will have a chance to present a case for Relationship (2) in the near future. So, applying either relationship we can see that the construct for the ordinals [def.art. + numeral+θ/ð] gets simplified in Tw to [def.art. + numeral] with the loss of the word-final morpheme for ordinal. To wit:
tē it 第一 (the first)
tē jī 第二 (the second)
tē sann 第三 (the third)
etc.
It is interesting to observe that, if we assume tē 第 as the vestige of the def.art., it is retained in the ordinals in analogous position, even though Tw has found no use for the def.art. generally.
Heruler
P.s. Sorry for the delay in responding to you. I just got back from a trip to Los Angeles and saw your comments. During my trip to the lovely Los Angeles, someone broke into my rental car and stole all my lugguage, including my notebook (not the laptop notebook, just the paper-based notebook) in which I jotted down my incidental findings in language correspondences. I pray that I will stumble upon those findings again.
Very happy to see that you have responded to my posting on the Taiwanese (Tw) ordinals. Your criticism is well founded in linguistics, and I felt your commentary has taken the forum to a new height.
Your penetrating questions truly touched the heart of the issue. I will try to respond to them one by one. Let me address the first question first. What you said is right about the ordinals in Germanic languages whose characteristic element is missing in Taiwanese. Let me explain the missing link in a slightly expanded fashion. The ordinals in Germanic languages are composed of a definite article (def.art.) and a numeral suffixed with a morpheme for the ordinal. To form the ordinals other than the first, second, and third, Indo-European (IE) has an affix * + t. This was retained in Germanic (Gmc.) in cases like *sexst+ ‘6th’, but generally by Verner’s law had changed to Gmc. * +θ or * +ð. (Reference: Joseph B. Voyles, 1992. Early Germanic Grammar. Pre-, Proto-, and Post-Germanic Languages, Academic Press, San Diego, CA. p. 248.) In Old Norse (ON), the affix is either -di or -ði. Only in a few cases is the affix -ti, e.g., sétti ‘6th’.
When I compared Tw words with Germanic in general, I found the following interesting relationships. (1) If the Gmc. words are monosyllabic and the words end in a consonant cluster, the consonants are either dropped or simplified in Tw words. (2) If the Gmc. words have more than 1 syllable, the first syllable is usually retained, and the rest is “lost”. Only occasionally do we find 2 syllables are retained, e.g., ON teina (basket made of twigs) : Tw tîn-nâ 藤藍 ; ON amma (grandmother) : Tw a-má ; ON orð-lof (praise) : Tw o-ló (praise).
In the next posting (in a new thread), I will present a pattern of sound correspondence between ON and Tw to illustrate Relationship (1). Hopefully I will have a chance to present a case for Relationship (2) in the near future. So, applying either relationship we can see that the construct for the ordinals [def.art. + numeral+θ/ð] gets simplified in Tw to [def.art. + numeral] with the loss of the word-final morpheme for ordinal. To wit:
tē it 第一 (the first)
tē jī 第二 (the second)
tē sann 第三 (the third)
etc.
It is interesting to observe that, if we assume tē 第 as the vestige of the def.art., it is retained in the ordinals in analogous position, even though Tw has found no use for the def.art. generally.
Heruler
P.s. Sorry for the delay in responding to you. I just got back from a trip to Los Angeles and saw your comments. During my trip to the lovely Los Angeles, someone broke into my rental car and stole all my lugguage, including my notebook (not the laptop notebook, just the paper-based notebook) in which I jotted down my incidental findings in language correspondences. I pray that I will stumble upon those findings again.
Hopefully both of you will keep posting more. Dear Heruler,
Sorry to hear about your bad experience in LA - any kind of theft is extremely annoying, hope you didn't lose too much! I have just come back from a business trip to Brazil, where a friend of mine was robbed at gunpoint (!) of both his car and laptop ... at least the robbers left him himself unharmed.
Having just come back from abroad is also the reason for my long delay in responding. I am happy to see that you have taken on my challenge and I'll be glad to keep this going - I must admit that I do not have much hope of finding a Germanic/ Norse - Chinese/ Minnan connection, but I do enjoy the intellectual exercise
I'll answer to the tentative -ang(r) correspondences in a proper separate e-mail - let me limit myself here to a few remarks on the numbers and a few general questions.
So we both agree on a suffix -t- being the principal formative element in IE to build ordinals, like in Latin sex (6) -> sex-t-us and German Sechs => der sechste. It's NOT that definite article that makes an ordinal number - after all, you can use 'die Sechs' in German if you want to talk about the (cardinal) number 6, like in playing cards or rolling dice .... Quite a few IE languages did not have a definite article in recorded history (e.g. Latin) - although that is a weaker argument, because it could always have been lost in some and retained in others ... but at the end of the day, it's not the definite article in European that makes the ordinal number but a suffix (unlike the Minnan 'te'). If 'te' in Minnan was really the definite article, this would be the only case where it was retained while the actual ordinal element was lost - not extremely plausible ...
The most important point for the Norse connection you are making is, however, that the Norse definite article is hinn or hit (depending on gender) - from where there is no phonetic way to 'te'
. Most Germanic languages base their article on a t-stem - but Norse doesn't!
Which brings me to the principal question: What exactly is the nature of the link you are trying to prove - is it a link between Germanic and Chinese? Between Indo-European and Chinese? Between Minnan and Norse specifically?
Let me explain why I think this question is important:
I hope we'll agree on the following statements:
(1) It's a proven fact that all Germanic languages come from a single (hypothetical) Common-Germanic ancestor.
(2) It's a proven fact that this common Germanic ancestor forms part of a larger IndoEuropean family, i.e. is descended from a common IE ancestor.
(3) It's a proven fact that Minnan, Yue, Mandarin, Wu etc. all are descended from a common Chinese ancestor.
(4) Minnan contains mainly Chinese ancestry but also some words that have no cognates in older literary Chinese and the modern fangyan and which are best explained as a non-Chinese substrate of the original inhabitants of the South before the Han immigration.
If we agree on all of these, how can we assume a Minnan-Norse connection? Let's see:
Model (1) - IE and Chinese are both descendants of a transcontinental superfamily. In that case, we should best look at reconstructed IE and Middle or Ancient Chinese to find correspondences, rather than looking at two descendents only (like Norse and Minnan - after all, there are a lot words which get dropped in one dialect and retained in another). Non-han words should not be included in the analysis because they were probably added to Minnan after the two families separated. You'd have to look at all IE languages and Chinese fangyan for valid comparisons - if something ties in Minnan but not in Yue, you'd need an explanation for the exception.
Model (2) - IE is descended from Chinese or vice versa - just a question of historic precedence - otherwise same conclusions as in (1).
Model (3) - There was language contact at some point in time which influenced certain word forms/ patterns in either Norse or Minnan - without requiring these to have influenced the other languages within their respective families - this model would be less strict - you could establish correspondences without having to look at all languages within the families. If, however, a word in Minnan shows a normal correspondence to the words in other fangyan and the supposed cognate in Norse developed normally in relationship to the other Germanic (or IE languages) then it's probably fair to assume that we do just have a case of coincidence but not of language contact (or of 'Urverwandschaft' like in model 1 or 2).
Please let me know which of these models you're actually trying to prove!
What I'm getting at is: If there is a seeming correspondence between two words in those two languages (Norse and Minnan) but no correspondence between the common IE ancestor and the common Chinese ancestor of this word, language contact is the only thing you can assume. If it's not language contact but direct descendence you're trying to prove, you'll have to be very stringent in checking the other family members, too.
Best regards,
Aurelio
Sorry to hear about your bad experience in LA - any kind of theft is extremely annoying, hope you didn't lose too much! I have just come back from a business trip to Brazil, where a friend of mine was robbed at gunpoint (!) of both his car and laptop ... at least the robbers left him himself unharmed.
Having just come back from abroad is also the reason for my long delay in responding. I am happy to see that you have taken on my challenge and I'll be glad to keep this going - I must admit that I do not have much hope of finding a Germanic/ Norse - Chinese/ Minnan connection, but I do enjoy the intellectual exercise
I'll answer to the tentative -ang(r) correspondences in a proper separate e-mail - let me limit myself here to a few remarks on the numbers and a few general questions.
So we both agree on a suffix -t- being the principal formative element in IE to build ordinals, like in Latin sex (6) -> sex-t-us and German Sechs => der sechste. It's NOT that definite article that makes an ordinal number - after all, you can use 'die Sechs' in German if you want to talk about the (cardinal) number 6, like in playing cards or rolling dice .... Quite a few IE languages did not have a definite article in recorded history (e.g. Latin) - although that is a weaker argument, because it could always have been lost in some and retained in others ... but at the end of the day, it's not the definite article in European that makes the ordinal number but a suffix (unlike the Minnan 'te'). If 'te' in Minnan was really the definite article, this would be the only case where it was retained while the actual ordinal element was lost - not extremely plausible ...
The most important point for the Norse connection you are making is, however, that the Norse definite article is hinn or hit (depending on gender) - from where there is no phonetic way to 'te'
. Most Germanic languages base their article on a t-stem - but Norse doesn't!Which brings me to the principal question: What exactly is the nature of the link you are trying to prove - is it a link between Germanic and Chinese? Between Indo-European and Chinese? Between Minnan and Norse specifically?
Let me explain why I think this question is important:
I hope we'll agree on the following statements:
(1) It's a proven fact that all Germanic languages come from a single (hypothetical) Common-Germanic ancestor.
(2) It's a proven fact that this common Germanic ancestor forms part of a larger IndoEuropean family, i.e. is descended from a common IE ancestor.
(3) It's a proven fact that Minnan, Yue, Mandarin, Wu etc. all are descended from a common Chinese ancestor.
(4) Minnan contains mainly Chinese ancestry but also some words that have no cognates in older literary Chinese and the modern fangyan and which are best explained as a non-Chinese substrate of the original inhabitants of the South before the Han immigration.
If we agree on all of these, how can we assume a Minnan-Norse connection? Let's see:
Model (1) - IE and Chinese are both descendants of a transcontinental superfamily. In that case, we should best look at reconstructed IE and Middle or Ancient Chinese to find correspondences, rather than looking at two descendents only (like Norse and Minnan - after all, there are a lot words which get dropped in one dialect and retained in another). Non-han words should not be included in the analysis because they were probably added to Minnan after the two families separated. You'd have to look at all IE languages and Chinese fangyan for valid comparisons - if something ties in Minnan but not in Yue, you'd need an explanation for the exception.
Model (2) - IE is descended from Chinese or vice versa - just a question of historic precedence - otherwise same conclusions as in (1).
Model (3) - There was language contact at some point in time which influenced certain word forms/ patterns in either Norse or Minnan - without requiring these to have influenced the other languages within their respective families - this model would be less strict - you could establish correspondences without having to look at all languages within the families. If, however, a word in Minnan shows a normal correspondence to the words in other fangyan and the supposed cognate in Norse developed normally in relationship to the other Germanic (or IE languages) then it's probably fair to assume that we do just have a case of coincidence but not of language contact (or of 'Urverwandschaft' like in model 1 or 2).
Please let me know which of these models you're actually trying to prove!
What I'm getting at is: If there is a seeming correspondence between two words in those two languages (Norse and Minnan) but no correspondence between the common IE ancestor and the common Chinese ancestor of this word, language contact is the only thing you can assume. If it's not language contact but direct descendence you're trying to prove, you'll have to be very stringent in checking the other family members, too.
Best regards,
Aurelio
Hi again,
I thought I should explain why I think it's a waste of time to try to find a connection between those languages:
If you look within the IndoEuropean languages, you'll find a lot of words which are obviously related across the entire continent. I'll give you a few examples:
(1) night - protogermanic nakht [naxt], Latin noct-, Russian [notsh'], Greek nuks, Sanskrit naktam ------ Mandarin ye4, Cantonese ye6, Minnan ia7 (or ammi).
(2) eight - protogermanic akhto, Latin octo, Greek okto, Sanskrit astau --- Mandarin ba1, Cantonese baat3, Minnan pat/ pueh
(3) head - protogermanic khaubuthan (=> old English heafod => head), Latin caput-, Sanskrit kaput-, --- Mandarin tou2, Cantonese tau4, Minnan tau5
& so forth
Looking at these examples (and there are many, many more), it becomes obvious that the IndoEuropean languages are closely related and that the Chinese languages are closely related among themselves - but it seems completely impossible to me to recognize any relationship between the two groups in those common words like numbers and basic everyday concepts.
Thus, the only thing you could try to prove would be language contact between say Minnan and Norse - however unlikely this may seem historically. If we are talking about Norse taking loan-words from Minnan or the other way round, this does require that e.g. a Minnan word in question cannot be explained in its Minnan form from a common Han ancestor - which leaves you with the rather small number of non-Han words in Minnan (like bah, kina, ti, sui, etc.).
While anybody should feel at liberty to seek connections between these and Norse (or any other language) it seems to me that they are much better explained as remnants of the SouthEast Asian languages spoken in Minnan territory before the Han immigration. This would make much more historic sense.
Likewise, if one assumes that certain words in Norse can only be explained as loanwords from Minnan, one has to show that they do not relate to other word in Germanic. The ang(r) examples given in the other post do relate to German quite nicely.
Best regards,
Aurelio
I thought I should explain why I think it's a waste of time to try to find a connection between those languages:
If you look within the IndoEuropean languages, you'll find a lot of words which are obviously related across the entire continent. I'll give you a few examples:
(1) night - protogermanic nakht [naxt], Latin noct-, Russian [notsh'], Greek nuks, Sanskrit naktam ------ Mandarin ye4, Cantonese ye6, Minnan ia7 (or ammi).
(2) eight - protogermanic akhto, Latin octo, Greek okto, Sanskrit astau --- Mandarin ba1, Cantonese baat3, Minnan pat/ pueh
(3) head - protogermanic khaubuthan (=> old English heafod => head), Latin caput-, Sanskrit kaput-, --- Mandarin tou2, Cantonese tau4, Minnan tau5
& so forth
Looking at these examples (and there are many, many more), it becomes obvious that the IndoEuropean languages are closely related and that the Chinese languages are closely related among themselves - but it seems completely impossible to me to recognize any relationship between the two groups in those common words like numbers and basic everyday concepts.
Thus, the only thing you could try to prove would be language contact between say Minnan and Norse - however unlikely this may seem historically. If we are talking about Norse taking loan-words from Minnan or the other way round, this does require that e.g. a Minnan word in question cannot be explained in its Minnan form from a common Han ancestor - which leaves you with the rather small number of non-Han words in Minnan (like bah, kina, ti, sui, etc.).
While anybody should feel at liberty to seek connections between these and Norse (or any other language) it seems to me that they are much better explained as remnants of the SouthEast Asian languages spoken in Minnan territory before the Han immigration. This would make much more historic sense.
Likewise, if one assumes that certain words in Norse can only be explained as loanwords from Minnan, one has to show that they do not relate to other word in Germanic. The ang(r) examples given in the other post do relate to German quite nicely.
Best regards,
Aurelio
Hi Aurelio,
Thanks for your comments. Sorry for the long delay in submitting my reply - I have been quite tied up in my work. You raised a few interesting points to which I would like to respond in small installments (due to my time constraints). Let me first address the most important issue regarding the model. For the following discussion, I would like to use the term Holó, rather than Taiwanese or Hokkien, because it refers to a wider area than just Hokkien and Taiwan; it covers an area extending from the southern coast of Zhejiang to northern Guangdong, pockets in Guangdong, to Hainan, and further to SE Asia and southern California.
The model I have in mind is the Norman Conquest of England in 1066. From Christmas Day, 1066, when William the Conqueror was crowned King of England, to 1362 when Parliament admitted that French was “much unknown in the said realm” (Stevenson, p. 136), French was the language of the English court and the aristocrats. So anybody whose mother tongue was the “lowly” English and aspired to climb up the social ladders had to learn French. Thousands of French words were introduced for government, church, law and the military. In the fourteenth century when English began to be on the ascendance, those who were bilingual spoke English mixed with French vocabulary. This trend added thousands more French words to the English vocabulary. So overall the main effect of Norman Conquest was introduction of an enormous number of French loan-words into English, which drove out native English words, making Old English texts so difficult for us to read now. The end result is that more than 60% of the Old English vocabulary have disappeared and replaced by French (Trask, p. 20; Barber, pp. 134-150).
There are many pairs of French and English words which reflect the master and servant, ruler and ruled relationships. The Old English calf, ox, pig and sheep of the Saxon farms became veal, beef, pork and mutton on the table of the Norman nobles. This aspect is nicely weaved into a conversation between two shepherds in the opening chapter of Sir Walter Scott’s Ivanhoe, the classic tale of the medieval knights of England.
Today the Normans have completely been assimilated in the English population. So, there is no separate group of Normans living in England now. Let’s imagine we board a time machine to transport us to the fifteenth century England. We take a leisure tour of the country and somehow wander into a remote county. We meet friendly residents who speak a peculiar language that, with your trained ear for a variety of languages, you have no difficulty in recognizing as Anglo-Norman mixed with Middle English. We surmise that they may be descendants of the Norman nobility who have left the aristocracies two centuries earlier and now live a quiet and isolated country life. Thus they are able to retain most of their native French vocabulary, although the language structure may resemble English.
Such is my model. To follow the model I make a couple of suppositions.
First supposition. I suspect Holó people are the raimnants of old noble classes or closely related to the royal house. About 70% of its vocabulary is shared with Chinese. The remaining 30% have no cognates in Chinese. (I will list the sources for my estimate in the next posting.) I designate the latter as Group-1 words (G1W), and the former the Group-2 words. The G1W are like the Anglo-Norman words that you heard in that imaginary county in England, which are not shared with English. Quite a number of Group-2 words in Chinese are used in very polite and formal situation only, suggesting that they are words from a language once considered prestigious.
Second supposition. I think that a people or a group of peoples migrated from Europe to Asia and ruled over China for some time (like the Norman ruler of England). Therefore, their vocabulary was borrowed by the native Chinese, while the rulers also picked up some Old Chinese vocabulary as well. Thus it can account for the 70% of its vocabulary sharing with Chinese. I do not know exactly how much of the 70% belongs to the Old Chinese category. On the Chinese side, I do not know how much of the Old Chinese vocabulary have been replaced by the loan-words. One fact is known, from Old Chinese to Middle Chinese the vocabulary expanded tremendously.
Throughout Chinese history there have been many invasions and conquests by foreigners who established dynasties. I suspect the massive introduction of foreign words occurred in the Northern Dynasties (北朝). The rulers of the Northern Wei Dynasty (as well as the Tang Dynasty) came from the Dai Region 代郡 (today’s northern Shanxi 山西), which in the early centuries AD was a gathering place of heterogeneous “peoples from outside the forts” 塞外民族. Their languages were vastly different from Chinese, but unfortunately there are no extant texts for us to study except for some scattered words. Perhaps more importantly, there are Chinese records (e.g., 顏氏家訓 Gân-sì Ka-hùn) showing that the native Chinese eagerly learned Hu-yu 胡語 (Barbarians' language) during the Northern Dynasties.
There is lexical evidence linking Holó, the current dialect of Shanxi, and the Japanese kan-on pronunciation of kanji, one of which is de-nasalization of certain words (Reference: Forest, Chinese Language, but I do not have the book at hand.) Japanese borrowed words from the Tang Court, thus its pronunciation is believed to be close to the Tang speech. The following words demonstrate the denasalization. Let me use Old Norse words and their Japanese and Holó corresponding words for comparison. (I do not have a dictionary of the Shanxi dialect so I do not know how they sound.) For Japanese words I use the Japanese spelling.
Old Norse marr (horse) : Japanese ba 馬 : Holó bé (Note: there is an a > e sound change.)
Old Norse mey (maid, girl) : Holó bē/bōe 妹
Old Norse mær (maid, girl) : Japanese bai 妹
Old Norse matr (meat) : Holó bah (G1W)
Old Norse niu (nine) : Holó kiú 九 : Japanese kyū
At some point the rulers lost power, or due to turmoils in northern China, they migrated south and found refuge in Fujian, where they were well protected by mountains in the rear. Their front yard is the ocean, and thus they spread out along trade routes to Taiwan, Hainan and onward to SE Asia. Being traders with good business acumen, they essentially made a Hanseatic League out of the South China Sea. The large geographic area, scattered with Holó communities, serves as a language reservoir, allowing the language to thrive. Anglo-Norman was not afforded such an opportunity to maintain a separate identity.
This is just a working hypothesis based on the Norman Conquest as the model. I know there is a lot of work to be done…
References:
Barber, Charles (1993). The English Language: A Historical Introduction. Cambridge University Press, Cambridge, UK. (ISBN 0-521-42622-7)
Stevenson, Victor (1999). The World of Words. Sterling Publishing Co., Inc., New York. (ISBN 0-8069-3905-2)
Trask, R.L. (1996). Historical Linguistics. Arnold/St. Martin’s Press, New York. (ISBN 0-340-60758-0)
Best regards,
Heruler
Thanks for your comments. Sorry for the long delay in submitting my reply - I have been quite tied up in my work. You raised a few interesting points to which I would like to respond in small installments (due to my time constraints). Let me first address the most important issue regarding the model. For the following discussion, I would like to use the term Holó, rather than Taiwanese or Hokkien, because it refers to a wider area than just Hokkien and Taiwan; it covers an area extending from the southern coast of Zhejiang to northern Guangdong, pockets in Guangdong, to Hainan, and further to SE Asia and southern California.
The model I have in mind is the Norman Conquest of England in 1066. From Christmas Day, 1066, when William the Conqueror was crowned King of England, to 1362 when Parliament admitted that French was “much unknown in the said realm” (Stevenson, p. 136), French was the language of the English court and the aristocrats. So anybody whose mother tongue was the “lowly” English and aspired to climb up the social ladders had to learn French. Thousands of French words were introduced for government, church, law and the military. In the fourteenth century when English began to be on the ascendance, those who were bilingual spoke English mixed with French vocabulary. This trend added thousands more French words to the English vocabulary. So overall the main effect of Norman Conquest was introduction of an enormous number of French loan-words into English, which drove out native English words, making Old English texts so difficult for us to read now. The end result is that more than 60% of the Old English vocabulary have disappeared and replaced by French (Trask, p. 20; Barber, pp. 134-150).
There are many pairs of French and English words which reflect the master and servant, ruler and ruled relationships. The Old English calf, ox, pig and sheep of the Saxon farms became veal, beef, pork and mutton on the table of the Norman nobles. This aspect is nicely weaved into a conversation between two shepherds in the opening chapter of Sir Walter Scott’s Ivanhoe, the classic tale of the medieval knights of England.
Today the Normans have completely been assimilated in the English population. So, there is no separate group of Normans living in England now. Let’s imagine we board a time machine to transport us to the fifteenth century England. We take a leisure tour of the country and somehow wander into a remote county. We meet friendly residents who speak a peculiar language that, with your trained ear for a variety of languages, you have no difficulty in recognizing as Anglo-Norman mixed with Middle English. We surmise that they may be descendants of the Norman nobility who have left the aristocracies two centuries earlier and now live a quiet and isolated country life. Thus they are able to retain most of their native French vocabulary, although the language structure may resemble English.
Such is my model. To follow the model I make a couple of suppositions.
First supposition. I suspect Holó people are the raimnants of old noble classes or closely related to the royal house. About 70% of its vocabulary is shared with Chinese. The remaining 30% have no cognates in Chinese. (I will list the sources for my estimate in the next posting.) I designate the latter as Group-1 words (G1W), and the former the Group-2 words. The G1W are like the Anglo-Norman words that you heard in that imaginary county in England, which are not shared with English. Quite a number of Group-2 words in Chinese are used in very polite and formal situation only, suggesting that they are words from a language once considered prestigious.
Second supposition. I think that a people or a group of peoples migrated from Europe to Asia and ruled over China for some time (like the Norman ruler of England). Therefore, their vocabulary was borrowed by the native Chinese, while the rulers also picked up some Old Chinese vocabulary as well. Thus it can account for the 70% of its vocabulary sharing with Chinese. I do not know exactly how much of the 70% belongs to the Old Chinese category. On the Chinese side, I do not know how much of the Old Chinese vocabulary have been replaced by the loan-words. One fact is known, from Old Chinese to Middle Chinese the vocabulary expanded tremendously.
Throughout Chinese history there have been many invasions and conquests by foreigners who established dynasties. I suspect the massive introduction of foreign words occurred in the Northern Dynasties (北朝). The rulers of the Northern Wei Dynasty (as well as the Tang Dynasty) came from the Dai Region 代郡 (today’s northern Shanxi 山西), which in the early centuries AD was a gathering place of heterogeneous “peoples from outside the forts” 塞外民族. Their languages were vastly different from Chinese, but unfortunately there are no extant texts for us to study except for some scattered words. Perhaps more importantly, there are Chinese records (e.g., 顏氏家訓 Gân-sì Ka-hùn) showing that the native Chinese eagerly learned Hu-yu 胡語 (Barbarians' language) during the Northern Dynasties.
There is lexical evidence linking Holó, the current dialect of Shanxi, and the Japanese kan-on pronunciation of kanji, one of which is de-nasalization of certain words (Reference: Forest, Chinese Language, but I do not have the book at hand.) Japanese borrowed words from the Tang Court, thus its pronunciation is believed to be close to the Tang speech. The following words demonstrate the denasalization. Let me use Old Norse words and their Japanese and Holó corresponding words for comparison. (I do not have a dictionary of the Shanxi dialect so I do not know how they sound.) For Japanese words I use the Japanese spelling.
Old Norse marr (horse) : Japanese ba 馬 : Holó bé (Note: there is an a > e sound change.)
Old Norse mey (maid, girl) : Holó bē/bōe 妹
Old Norse mær (maid, girl) : Japanese bai 妹
Old Norse matr (meat) : Holó bah (G1W)
Old Norse niu (nine) : Holó kiú 九 : Japanese kyū
At some point the rulers lost power, or due to turmoils in northern China, they migrated south and found refuge in Fujian, where they were well protected by mountains in the rear. Their front yard is the ocean, and thus they spread out along trade routes to Taiwan, Hainan and onward to SE Asia. Being traders with good business acumen, they essentially made a Hanseatic League out of the South China Sea. The large geographic area, scattered with Holó communities, serves as a language reservoir, allowing the language to thrive. Anglo-Norman was not afforded such an opportunity to maintain a separate identity.
This is just a working hypothesis based on the Norman Conquest as the model. I know there is a lot of work to be done…
References:
Barber, Charles (1993). The English Language: A Historical Introduction. Cambridge University Press, Cambridge, UK. (ISBN 0-521-42622-7)
Stevenson, Victor (1999). The World of Words. Sterling Publishing Co., Inc., New York. (ISBN 0-8069-3905-2)
Trask, R.L. (1996). Historical Linguistics. Arnold/St. Martin’s Press, New York. (ISBN 0-340-60758-0)
Best regards,
Heruler
Estimates of the percentage of Group-1 words in the Holó vocabulary:
(1) Carstairs Douglas (1873). Chinese-English Dictionary of the Vernacular or Spoken Language of Amoy, Trübner & Co., London. On page viii, he gave an estimate of 25 – 33%. This was based on his impression obtained when compiling his monumental dictionary.
(2) Ong Iok-tek 王育德 (2000). Tâi-oân-ōe Káng-chō 台灣話講座 (Collection of a series of papers published in 1960 – 1964), Avangard Publishing Company 前衛出版社, Taipei. On page 93, he compared 200 basic words (Swadesh 200) between Mandarin Chinese and Taiwanese, and found that 49% of the 200 words show lack of correspondence. Because the 200 basic words pertain mostly to family life, the divergence between Taiwanese and Chinese tends to be larger than the overall vocabulary, skewing the statistics.
(3) Robert Cheng 鄭良偉 (1981). “Borrowing and internal development – Comparison of Taiwanese words and their Mandarin equivalents.” Paper presented at the 14th International Conference on Sino-Tibetan Languages and Linguistics, Gainesville, Florida. (Cited in John DeFrancis, The Chinese Language, Fact and Fantasy, University of Hawaii Press, Honolulu, 1984, page 63). Professor Cheng gave an estimate of 30%.
Professor Cheng used modern linguistic analysis and his estimate of 30% falls within the range of 25-33% given by Douglas. Therefore, I adopt the 30% estimate.
Heruler
(1) Carstairs Douglas (1873). Chinese-English Dictionary of the Vernacular or Spoken Language of Amoy, Trübner & Co., London. On page viii, he gave an estimate of 25 – 33%. This was based on his impression obtained when compiling his monumental dictionary.
(2) Ong Iok-tek 王育德 (2000). Tâi-oân-ōe Káng-chō 台灣話講座 (Collection of a series of papers published in 1960 – 1964), Avangard Publishing Company 前衛出版社, Taipei. On page 93, he compared 200 basic words (Swadesh 200) between Mandarin Chinese and Taiwanese, and found that 49% of the 200 words show lack of correspondence. Because the 200 basic words pertain mostly to family life, the divergence between Taiwanese and Chinese tends to be larger than the overall vocabulary, skewing the statistics.
(3) Robert Cheng 鄭良偉 (1981). “Borrowing and internal development – Comparison of Taiwanese words and their Mandarin equivalents.” Paper presented at the 14th International Conference on Sino-Tibetan Languages and Linguistics, Gainesville, Florida. (Cited in John DeFrancis, The Chinese Language, Fact and Fantasy, University of Hawaii Press, Honolulu, 1984, page 63). Professor Cheng gave an estimate of 30%.
Professor Cheng used modern linguistic analysis and his estimate of 30% falls within the range of 25-33% given by Douglas. Therefore, I adopt the 30% estimate.
Heruler