10

In another Linguistics.SE question, an answer makes this claim:

The difference between French and German is so much larger than between Mandarin/Cantonese that one would be hard pressed to say they are of similar distance.

This however contradicts what I find with some very brief googling of the phrase 'lexical similarity':

Wikipedia states that the 'lexical similarity' between French and German is 29%.

While someone else claims that:

According to lexical statistical data of any two languages within the Sinitic branch (for example Wu and Mandarin), the data will always reveal that there is less intelligibility between them than any two Romance languages in Europe. For example, French has lexical similarity of about 75% to several other Romance languages. In comparison, Mandarin has 31% lexical similarity with Wu (Shanghainese) and 19% with Yue (Cantonese). Source

My question is therefore this:

Is it obvious that the difference between French/German is much larger than between Mandarin/Cantonese?

(Indeed, if the numbers in the above sources are to be believed, English is more similar to Russian [24%] than Mandarin is to Cantonese [19%]. Is this possible?)

5
  • 1
    First you say that Mandarin has a 19% "lexical similarity" with Cantonese, but five lines later you say that Mandarin is 19% "different" from Cantonese. Are you sure you know what you are saying?
    – fdb
    Commented Jul 22, 2014 at 13:30
  • 3
    These are not fixed figures. These are estimates, compiled from variable and incomplete data, with varying definitions of what "similarity" means. It is not obvious, in other words. There are no reliable statistics on language similarity in the abstract -- individual variables swamp everything here.
    – jlawler
    Commented Jul 22, 2014 at 17:12
  • PS. The mistake to which I drew attention has been (tacitly) corrected. Otherwise, I agree fully with jlawler.
    – fdb
    Commented Jul 23, 2014 at 9:31
  • 2
    Lexical similarity is just one measure of language difference though. The lexical similarity between English and French, for example, is likely to be higher than the structural, morphological or phonetic similarities. Certainly, the split between Southern Sinitic and Mandarin is more recent than that between Germanic and Italic.
    – Uri Granta
    Commented Mar 12, 2015 at 11:00
  • I can put together a lexical distance calculation between the two pairs for a group of words and disregarding any other language features. Commented Jun 3, 2017 at 8:43

3 Answers 3

4

No, it's not 'obvious' when you take into account both the lexical distance, and date of divergence.

I don't speak Mandarin or Cantonese, and I feel like it would be good to hear from a historical linguist who speaks both, but seeing as no one has answered this yet...

Historically, Germanic and Romance languages obviously split much earlier than Mandarin and Cantonese did, but the problem with that is that languages change at different rates.

There is not a widely-accepted way of measuring language distance in historical linguistics. There are some proposals like Glottochronology, but the problem with Glottochronology, is again, that languages change at different rates.

If we think about this in the framework of evolutionary biology, then Mandarin and Cantonese probably last shared a common ancestor in early Middle Chinese, spoken around 600 AD, with a rhyme table recorded in the Qieyun in 601.

Since then, Mandarin underwent many radical changes, including the loss of most word final consonants, which in turn created lots of homophones, which in turn caused speakers to make lots of compounds to disambiguate these homophones causing huge lexical changes.

Asking whether or not Mandarin is as lexically distant from Cantonese as English is to French elaborates only one dimension of the picture. Mandarin and Cantonese share a closer recent common ancestor than English and French do.

However, Mandarin and Cantonese are probably much more divergent than say, the Germanic languages, which also began diverging around that time.

I remember one textbook, Anthropology of Language, claiming Mandarin and Cantonese are as lexically divergent as English and Italian. I don't think it cited any sources, and I don't have the book on me.

It is important to keep in mind that Mandarin and Cantonese may be as divergent as English is to Italian, but that this change happened fast with Mandarin; within the last 1,400 years, while English and Italian probably last shared a common ancestor more than 3,000 to 5,000 years ago.

2

The lexical similarity is usually calculated using a Swadesh list of 100 or 200 core vocabulary and counting shared cognates. The lexical similarity of French and other Romance languages is estimated by Dyen et al. (1992) and Rea (1958) to be around 70–80%. French-German is around 20–30% by Dyen et al. (1992). That of Mandarin and Cantonese is estimated to be about 70% by Wang (1960) and Xu (1991) using a 200-word and 100-word list respectively. The diversity of Sinitic languages is comparable to that of Romance, although the former is somewhat more diverse but not significantly (Maxwell 2024). Also, here is a Swadesh list of Sinitic languages, and it is obvious that most of them share at least half the vocabulary.

The figure of 19% and 31% for Mandarin-Cantonese and Mandarin-Wu is way too low and is probably not an estimation of lexical similarity in this sense. A similar figure of 24% and 29% comes from Cheng (1982), which calculates a correlation index using a much longer wordlist specifically designed for collecting dialect vocabulary and takes multiple variants into account. When varieties share all the variants, the score is one; when they share none, the score is negative one. If they share some variants, the score ranges between negative and positive one. Thus, this is a correlation index rather than a measure of lexical similarity. Lexical similarity typically uses only the most common variant for each lexical item and counts cognates. In contrast, the correlation index in this study considers multiple variants. In lexical similarity, a score of zero would be given if no cognates are shared, rather than a negative score. Additionally, this method includes non-core vocabulary, which can skew the results. While this is useful for dialectology, it is not a measure of lexicostatistics. In summary, the figures provided in the question are derived from different estimation methods and cannot be directly compared.

Due to sampling error, for a 200-word Swadesh list a 99.7% confidence interval would be ±9.7% for Cantonese-Mandarin and ±9.1% for French-German (Maxwell 2024). After taking these into account, the difference between French/German is significantly larger than between Mandarin/Cantonese at least from a lexical perspective.

References:

  • Dyen, Isidore, Kruskal, Joseph B., Black, Paul (1992). An Indoeuropean Classification: A Lexicostatistical Experiment. Philadelphia: American Philosophical Society.
  • Rea, John (1958). Concerning the validity of lexicostatistics. International Journal of American Linguistics 24(2), pp. 145–150.
  • Wang, Yude (1960). Chugokugo dai hogen no bunretsu nendai no gengo nendai gakuteki shitan 中国五大方言の分裂年代の言語年代学的試探 [The lexicostatistic estimation of the time depths of the five main Chinese dialects]. Gengo Kenkyu: Journal of the Linguistic Society of Japan 38, pp. 33–105.
  • Xu, Tonquiang (1991). Lishi Yuyanxue 历史语言学 [Historical linguistics]. Beijing: Shangwu Yinshuguan.
  • Maxwell, A., McMillan, L. (2024). Error bars for lexicostatistical estimates, with a case study comparing the diversity of Chinese and Romance. Linguistica Brunensia 72(1), pp. 1–17.
  • Cheng, C. C. (1982). A quantification of Chinese dialect affinity. Studies in the Linguistic Sciences 12(1), pp. 29–47.
New contributor
Simonster is a new contributor to this site. Take care in asking for clarification, commenting, and answering. Check out our Code of Conduct.
1

Unlike Ghoti657, I do speak Cantonese and some Mandarin, but I speak neither French nor German.

I can offer my opinion that German, being a Germanic language, and French, being a Romance language are quite different because of spelling differences and differences in phonology.

The difference between Cantonese and Mandarin, however, is different depending on what you are analyzing. The writing system and vocabulary is largely identical even after accounting for the fact that mainland China uses simplified characters. Common (like pronouns) and colloquial words will be more divergent, but Cantonese does not have too many dialectical characters of its own and the vocabulary is fairly close due to a shared literary tradition, and being nearly directly descended from the same historical language. However, written Cantonese may still be unintelligible depending on the subject. Where the subject matter is more technical, the vocabulary will tend to be very similar like English use of Latin-derived words for this purpose.

Spoken, there is a large difference between the two with the main difference being that Cantonese stop codas are absent in Mandarin, and the rich Mandarin vowel sequences are not present in Cantonese. There is enough correspondence between Cantonese and Mandarin phonology that sometimes the pronunciation of the other language can be guessed with some training and a good command of one's own native language. Multi-syllable words can sometimes be understood between the two and the grammar is comparable, although I have trouble with standard Mandarin Chinese grammar.

So spoken, I would liken the difference between Cantonese and Mandarin, then, as between Italian and French. Written, I would liken it to the difference between Spanish and Italian.

German and French, though, are like Mandarin and Fujianese (which hypothetically split off before Middle Chinese even came into existence).

1
  • I speak both Cantonese and Mandarin, know very little German and even less French. I have no training or indeed any learning in linguistics whatsoever, which is why I posted my question here. I would clarify though that although traditionally, Cantonese simply has the same shared written tradition as Mandarin, in fact many Cantonese words are completely different and unknown in Mandarin. This is shown in Hong Kong, where in the past 10+ years or so there've increasingly been Cantonese words in print (mostly to defy China and also to maintain a distinct identity).
    – user3222
    Commented Jul 9, 2017 at 23:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.