朱学渊按：大家或许知道我妻子四年前得了很严重的癌症，去年五月去世，我是在这以后才重新开始工作。此后接到数百封支持和鼓励信函，但我只发表那些反对我的意见，因为反对意见能够激发我思考和写作的热情。有一位朝鲜族读者来信说，他从《百度》上读到我的母亲是朝鲜族，因此他很高兴；虽然我母亲是浙江温州市瑞安县人，但是我还是为能被民族同胞认同而高兴，于是我去查阅一下。见内容非但如此，后面还有姚大力先生的批评，如“朱学渊……缺乏语言学和语音学的相关知识”等等（http://baike.baidu.com/view/80113.htm）。大约十五年前，我与姚大力先生有一面同席之识，后来他在《文汇报》责备中华书局发出版鄙著《中国北方民族的源流》是“学术错误”，我与中华书局柴剑虹先生也曾与他争论。十年前他又在《往复网》引称他是得到余英时先生的支持云云（我已忘记原话）。众所周知也完全可以理解，普林斯顿大学荣誉退休教授余英时先生对别人（包括毛泽东蒋介石）的指责常常是“边缘人物”，在他的心目中唯长青藤学校的饭碗才是“正当职业”，除《朱熹哲学》是中华主流，余皆“旁门左道”（“边缘”之别写）。去年我曾去台湾一次，得知他曾以某种方式暗示不要发表我的文章，被示者则将那些文章结集成书。其实这些事情我早以明晰，也并不在乎。而是姚大力先生披着虎皮，说我这也“不懂”那也“不懂”，还要改我的宗，叫我十分生气。我学量子物理无所作为，却来抢历史主流内行的饭碗，的确可以被他们认为是无道德的边缘人物；但是，如果我以姚先生从工农兵直升研究生，而指责他是王洪文式的“政治骗子”又妥当吗？事实上，我非但不以“边缘”为耻辱，反以为“边缘”是突破的“机缘”。今年我已经七十三岁，余先生必八十有多，姚先生也不年轻，即便有再高的资历和荣耀，大家还须多做实事，多立新功；如果带着过于骄傲自负的判官心态升天，上帝或许会指责那是有违学术探索者应有之品格的，此见。二○一五年七月二十五日
又及：后附两位语言学家半年前的来信，本邮件系统不许随附附件，许毅先生的英文文章中图片也无法贴上，而且内容我也没有读懂，但是我想本读者群中必有高明。还望识姚余二大师者转达本文。
伦敦大学学院（http://baike.baidu.com/view/16434.htm）语言学教授许毅教授来函：
学渊兄，“汉语是在戎狄语言的基础上发生的较年轻的语言”，是很有意思的断言。我们目前对韵律的研究已有初步证据印证这一假说。见附文。许毅
中国民族语言学大师，藏语专家瞿霭堂教授来信：
学渊先生，拜读这次来文，觉得道理是通的，对汉语汉文的产生和发展有启发意义。古代中原地区，的确存在阿尔泰语系的语言，而且占据重要地位。汉语是多种民族语言混合产生的，一般认为到周朝的雅言才定型。当然，周以前早就有汉语，起始时期和发展过程要比周朝早得多，甲骨文写的已经是汉语，也就是说，至少殷商时期汉语已经形成。汉语什么时候产生，无法考证。由于使用汉语的人与使用阿尔泰语言的人在中原杂居，语言之间互相渗透，彼此语言之间都有对方的成分。汉族形成过程中有说阿尔泰语言的人掺杂起来，也没有问题的。汉族是多种民族融合形成的。我著文论述了藏缅语族的语言，是汉语和阿尔泰语混合而成的，既不是汉语，也不是阿尔泰语言，是混合的语言，他们在语言上没有共同的祖语。也就是说，藏缅语族语言与汉语不是由一种语言发展变化而来的。这点是与你来文中说法不一样的（你认为“与藏缅语同类的汉语”，意思是同类的，我认为是不同类的）。
历史语言学还没有很好解决一个问题：一种语言（A）中有其他语言（B）的成分，这种成分是其他民族（B）学习这种语言（A）时保留下来的（术语叫底层），还是这种语言（A）从其他语言（B）借入的。因此，不能因为A语言中有B语言的东西，就推断A语言是B语言变来的，或者使用A语言的人曾经先是使用B语言的。《尚书》和《逸周书》中有蒙古语成分很正常，说明这两个民族历史上有过密切接触，语言之间互相渗透过。也可以推论汉族形成过程中，有蒙古族成员加入过。但你的结论“中国历史并不讳言开创华夏的夏、周、秦是戎狄部落。”恐怕不行。夏、周、秦朝已经说的是汉语，无论他们曾经是什么部落来源，他们使用汉语，说明已经为使用汉语的部落同化，接受了使用汉语部落的文化，只是在使用汉语的人的文化中渗透了一些他们原来部落的文化。即使使用汉语的部落的人使用汉字记录了他们的一些文化，在使用汉语的人的文化中也只是一小部分。因为你没有证明，使用汉语的部落起源于使用阿尔泰语言的部落。使用甲骨文书写汉语的部落是独立地形成的，其中可以包括阿尔泰成员，但是一种独立的语言和独立的文化，与使用阿尔泰语的人的部落和语言没有继承关系。你说“三、四千年前，黄河流域发生了语言的转型，与藏缅语同类的汉语取代了今天被归为“阿尔泰语系”的戎狄语言。”这里逻辑上和语言理论上有问题的。语言是怎么转型的？是使用阿尔泰语言的人自动转型的？这是不可能的。阿尔泰语言凭借什么道理自己发生这种大转型？（阿尔泰语是sov型，汉语是svo型）你能说明吗？转型必定是一个说汉语的部落战胜了说阿尔泰语言的部落，同化了他们的语言和文化。可见，夏、周、秦即使起源于戎狄部落，这时早已为使用汉语的部落所同化，否则如何解释一个使用阿尔泰语言的人会使用甲骨文书写另外一种语言？不写自己的语言？甲骨文文化不是阿尔泰文化，是汉族和汉语的文化。说夏、周、秦来源于戎狄部落可以，但当时已经为说汉语和写甲骨文的部落同化掉了。说汉语写甲骨文的部落是独立的部落，无论他们形成过程中有多少种部落掺合其中。这个部落是汉族和说汉语的祖先，与说阿尔泰语的部落没有继承或转变的关系。
区区陋见，望勿见笑。瞿霭堂
%%%%%%%%%%%%%%%%%%
附许毅先生论文
ICPhS XVII Special Session Hong Kong, 17-21 August 2011
POST-FOCUS COMPRESSION: CROSS-LINGUISTIC DISTRIBUTION AND HISTORICAL ORIGIN
Yi Xu
University College London, UK
ABSTRACT
One of the most important acoustic correlates of prosodic focus is post-focus compression (PFC) — the reduction of pitch range and amplitude of all post-focus components in an utterance. PFC has been found in many Indo-European, Altaic languages, and interestingly, also in Mandarin Chinese. Meanwhile, there have also been reports that many other languages do not have PFC, or lack any clear prosodic marking of focus. This paper presents a brief review of the current state of the art in the investigation of PFC, and discusses a number of hypotheses in regard to this typological division among the world languages. In particular, the idea is explored that the distribution of PFC is related to the historical development of the world languages.
Keywords: prosodic focus, post-focus compression, PFC
1. INTRODUCTION
Prosodic focus has long been treated, explicitly or implicitly, as a core component of speech prosody. The traditional English school of intonation is almost entirely built around the notion of the nuclear tone, which is defined as the most prominent component of a sentence [25, 26]. The nuclear tone notion is also adopted into the AM theory of intonation, although no fundamental differences between the nuclear and pre-nuclear tones are assumed by the theory [30]. A major characterization of the nuclear tone is that it is the last major intonational movement in an utterance, and subsequent movements, known as the “tail” in the British tradition, are much reduced in magnitude. In the AM theory, the post-nuclear components are described as “deaccented”, because they lack apparent F0 movements. The present paper is mainly about the phenomenon of the tail or deaccenting, which will hitherto be referred to as post-focus compression (PFC), for reasons that will become clear later.
Figure 1: Time-normalized mean F0 contours produced by 4 speaker groups. Each curve is an average of 40 repetitions by 8 speakers. The vertical lines indicate syllable boundaries. The solid thin lines are from the no-focus condition.
a) 50 100 150 200 250 Normalized time F0 (Hz) Initial Focus Medial Focus Final Focus Monolingual Beijing Mandarin
b) 50 100 150 200 250 Normalized time F0 (Hz) Initial Focus Medial Focus Final Focus Monolingual Taiwanese
c) 50 100 150 200 250 Normalized time F0 (Hz) Initial Focus Medial Focus Final Focus Monolingual Taiwan Mandarin
In the nuclear tone concept, the primary concern is given to the F0 contour of the focused word, especially that of the stressed syllable, while what happens in the tail is viewed as less important. This is probably because the large pitch movements around the nucleus is more easily noticeable when describing intonational forms without systematically controlling their functional relevance. In experimental research, the question-answer paradigm has been developed to systematically control the presence as well as the location of what is now widely known as the focus of the sentence, and to examine the acoustic patterns that co-vary with focus. In this paradigm, the focused component of a sentence directly corresponds to what is asked in the question. Using the question-answer paradigm, it is found that in languages as diverse as Swedish [4], American English [10, 40] and Mandarin [39], the pitch range of the focused word is raised and expanded, that of the post-focus words substantially lowered ICPhS XVII Special Session Hong Kong, 17-21 August 2011
and narrowed, while that of the pre-focus words largely unchanged. An example is shown in Figure 1a in which mean F0 contours of the Mandarin sentence “mama mo maomi” [Mom strokes kitty] spoken with focus on the first, second, third or none of the words are overlaid in the same graph. The findings of these studies suggest that that post-focus lowering of F0 is just as consistent as F0 raising on the focused word, suggesting at least equal importance of the post-focus “tail” as the on-focus F0 movement. Also from Figure 1a it can be seen that the largest F0 movement actually consists of mostly a sharp drop from the on-focus component to the first post-focus component. Thus much of the nuclear tone is actually the F0 movement of PFC.
The importance of post-focus pitch compression is further demonstrated by findings about focus perception. It is shown that a non-final focus can be perceived only when later occurring F0 peaks are very small, otherwise listeners would hear an additional late focus or no focus [24, 32]. Furthermore, focus recognition is much better when PFC is applicable, i.e., when focus is not sentence-final, whereas sentence-final focus, for which PFC cannot apply, is often perceptually confused with no focus [3, 7, 23, 32].
2. NON-UNIVERSALITY OF PFC
Prosodic patterns suggestive of PFC has been reported for many other languages, including Dutch [32], Greek [3], French [11], Korean [22], Turkish [17], Nanchang [36], German [13], Japanese [18], Uyghur, Tibetan [36], Arabic [5, 16], Hindi [28], Persian [33] and Finnish [24]. For all these languages, there is observation of post-focus F0 lowering, deaccenting or pitch range compression. It may thus seem that PFC is quite wide spread across languages. However, there is also evidence that in many other languages PFC is absent. For example, Taiwanese and Cantonese, both Chinese languages closely related to Mandarin, are found to lack PFC [27, 38]. As can be seen in Figure 1b, the F0 contours of the same sentences as in Figure 1a but spoken in Taiwanese show no sign of PFC in F0, and in fact very little difference can be seen across the four focus conditions. Also no PFC in intensity was found for either Taiwanese or Cantonese [7, 38]. Evidence of lack of PFC is also found for many other languages, including Yucatec Maya [19], Chichewa [41], Chitumbuka [41], Durban Zulu [41], Hausa [41], Yi, Deang, Wa [36], Buli [41], Northern Sotho [41] and Wolof [31]. Thus it is clear that, despite its apparent benefit for focus perception [3, 7, 23, 32], PFC is not universal.
3. ORIGIN OF PFC
3.1. Possible linguistic factors
There are many conceivable factors that could determine the present of PFC in a language, but some of them can be already ruled out. First, all languages in the Chinese family are fully tonal, and thus the fact that PFC can be either present or absent across these languages means that lexical tone cannot be the determining factor for PFC. Second, PFC might have to do with the fact that Mandarin has the neutral tone, which shares many properties with unstressed syllables in languages like English and German [40], while Taiwanese and Cantonese both lack equivalent of the neutral tone. However, there are also PFC languages that lack lexical stress, such as Japanese, Korean, French and Yi. Thirdly, the Chinese languages examined so far all have morpho-syntactic means to indicate focus. In fact it is claimed that morpho-syntactic focus markers exist in various forms in every language that has been examined [41], and therefore they are unlikely to be responsible for PFC. Of course, these considerations are not the final word on these factors, as new light could be shed by further research.
3.2. Origin of PFC in Mandarin（官话）
One of the most intriguing findings of Chen, et al. [7] is that Taiwan Mandarin, just like Taiwanese, also lacks PFC, as can be seen in Figure 1c. Taiwan Mandarin is a variant of Mandarin spoken in Taiwan. Although once homogeneous with Standard Chinese, at least by definition, it now has noticeable differences in vocabulary, grammar [8] and pronunciation [14] from its mainland counterpart. Today, most people in Taiwan are bilinguals, fluent in both Taiwanese and Taiwan Mandarin. Thus it is possible that Taiwan Mandarin has lost PFC due to close contact with Taiwanese (or with other southern Chinese dialects as well) through pervasive bilingualism in Taiwan over several generations.
If this is the case, a natural question would be, how did PFC occur in Mandarin in the first place, given that it is absent from so many other tone languages? There could be at least three possible hypotheses: (a) independent genesis, i.e., emerging ICPhS XVII Special Session Hong Kong, 17-21 August 2011
locally in the language, (b) horizontal spreading, i.e., borrowed into the language through contact, and (c) vertical inheritance, i.e., passed on from a proto-language. Regarding local emergence, the fact that PFC did not arise automatically in so many other languages at least suggests that it might not easily emerge in a language. In regard to spreading, a further question would be, from which language could PFC have been spread into Mandarin?
Historically, northern China was in close contact with many non-Chinese speaking populations, in particular, Mongolian and Manchurian, who ruled China during the Yuan (1271-1368 AD) and Qing (1644-1912 AD) dynasties. As a result, there has been much influence of those languages on Mandarin [6, 21, 35]. Both Mongolian and Manchurian are Altaic, a hypothetic language family that includes Turkic, Mongolic and Tungusic languages [15], and Korean and Japanese according to some scholars [15]. Interestingly, all these languages have shown evidence of PFC [18, 17, 22, 36]. Thus it is possible that PFC in Mandarin came from Altaic languages via contact.
The difficulty with the spreading account, however, is that so far there is evidence only for the loss of PFC through language contact, but no case of a language gaining PFC through contact. This is true of native Mandarin-Taiwanese bilinguals [7], English-Cantonese bilinguals [37], Sotho speakers learning English as a second language [34], Cantonese [37], Deang, Wa and Yi [36] speakers learning Mandarin as a second language. What is not yet known is the case of a PFC language speakers learning a non-PFC language as a second language. Would these learners at least initially carry PFC into their non-proficient second language?
If it is indeed not easy for PFC to spread or automatically emerge, the inheritance hypothesis may have to be taken seriously. The implications of this hypothesis would be profound, however. First, it would mean that Mandarin is a descendant of an Altaic language, but has acquired, through language contact, a large amount of characteristics shared with other Chinese languages, including tone. This scenario would challenge the current assumption that all Chinese languages are derived from a single proto-Chinese language [1]. Interestingly, there is already evidence from population genetics that the southern and northern populations in China are actually quite divided, suggesting rather different hereditary routes [9].
Secondly, vertical inheritance would mean that all cases of PFC can be traced back to an ancestral language where PFC first emerged, That is, there is a common ancestor to all modern PFC languages. This proto-language would have to be very ancient. From the distribution pattern that is currently emerging, the grouping of the PFC languages seem to be consistent with the hypothetical Nostratic superfamilty, consisting of the Indo-European, Uralic, Altaic, Afroasiatic, Dravidian, Kartvelian and Eskimo-Aleut language families [2, 29]. Their common ancestor, the proto-Nostratic, could be dated back to the end of the last Ice Age, i.e., 15,000-12,000 BC, which was probably spoken along the Fertile Crescent [2]. However, three of the Nostratic families mentioned above — Dravidian, Kartvelian and Eskimo-Aleut languages —have not yet been studied for PFC. New research is therefore needed.
4. CONCLUSION
A brief overview of the empirical findings about prosodic focus has shown that PFC, though an effective means of conveying focus through prosody, is present only in some of the world languages. The distribution of PFC suggests that its presence in a language is largely independent of factors such as lexical tone, lexical stress and availability of morphosyntactic markers of focus. The finding of its unspreadability through language contact suggests that PFC is most likely passed on to a language through inheritance rather than spreading. These findings may have profound implications for language typology, historical linguistics and human evolutionary history. There is therefore a need for increased collaborative research in all these directions.
5. REFERENCES
[1] Bodman, N.C. 1980. Proto-Chinese and Sino-Tibetan. In Coetsem, F.V., Waugh, L. (eds.), Contributions to Historical Linguistics. Leiden: Brill.
[2] Bomhard, A.R. 2008. Reconstructing Proto-Nostratic: Comparative Phonology, Morphology, and Vocabulary. Leiden: Brill.
[3] Botinis, A., Fourakis, M., Gawronska, B. 1999. Focus identification in English, Greek and Swedish. Proceedings of the 14th International Congress of Phonetic Sciences San Francisco. 1557-1560.
[4] Bruce, G. 1982. Developing the Swedish intonation model. Lund University, Dept. of Linguistics Working Papers 22, 51-116.
ICPhS XVII Special Session Hong Kong, 17-21 August 2011
[5] Chahal, D. 2003. Phonetic cues to prominence in Lebanese Arabic. Proceedings of the 15th International Congress of Phonetic Sciences Barcelona, 2067-2070.
[6] Chappell, H. 2001. Language contact and areal diffusion in Sinitic languages: Problems for typology and genetic affiliation. In Aikhenvald, A.Y., Dixon, R.M.W (eds.), Areal Diffusion and Genetic Inheritance: Problems in Comparative Linguistics. 167-194.
[7] Chen, S.-W., Wang, B., Xu, Y. 2009. Closely related languages, different ways of realizing focus. Proceedings of Interspeech Brighton, UK, 1007-1010.
[8] Cheng, R.L. 1985. A comparison of Taiwanese, Taiwan Mandarin, and Peking Mandarin. Language 61, 352-377.
[9] Chu, J.Y., Huang, W., et al. 1998. Genetic relationship of populations in China. Proceedings of the National Academy of Sciences 95(20), 11763-11768.
[10] Cooper, W.E., Eady, S.J., Mueller, P.R. 1985. Acoustical aspects of contrastive stress in question-answer contexts. Journal of the Acoustical Society of America 77, 2142-2156.
[11] Dohen, M., Loevenbruck, H. 2004. Pre-focal rephrasing, focal enhancement and post-focal deaccentuation in French. Proceedings of the 8th International Conference on Spoken Language Processing Jeju, Korea, 1313-1316.
[12] Downing, L.J. 2008. Focus and prominence in Chichewa, Chitumbuka and Durban Zulu. ZAS Papers in Linguistics (ZASPiL) 49, 47-65.
[13] Féry, C., Kügler, F. 2008. Pitch accent scaling on given, new and focused constituents in German. Journal of Phonetics 36(4), 680-703.
[14] Fon, J., Chiang, W.Y., Cheung, C. 2004. Production and perception of two dipping tones (T2 and T3) in Taiwan Mandarin. Journal of Chinese Linguistics 32, 249-280.
[15] Georg, S., Michalove, P.A., Ramer, A.M., Sidwell, P.J. 1999. Telling general linguists about Altaic. Journal of Linguistics 35(1), 65-98.
[16] Hellmuth, S. 2006. Focus-related pitch range manipulation (and peak alignment effects) in Egyptian Arabic. Proceedings of Speech Prosody 2006 Dresden, Germany, PS4-12-164.
[17] Ipek, C. 2011. Phonetic realization of focus with no on-focus pitch range expansion in Turkish. Proceedings of the 17th International Congress of Phonetic Sciences Hong Kong.
[18] Ishihara, S. 2002. Syntax-phonology interface of Wh-Constructions in Japanese. Proceedings of Tokyo Conference on Psycholinguistics (TCP 2002) Tokyo, 165-189.
[19] Kügler, F., Skopeteas, S. 2007. On the universality of prosodic reflexes of contrast: The case of Yucatec Maya. Proceedings of the 16th International Congress of Phonetic Sciences Saarbrücken, Germany.
[20] Ladd, D.R. 2008. Intonational Phonology. Cambridge: Cambridge University Press.
[21] LaPolla, R.J. 2001. The role of migration and language contact in the development of the Sino-Tibetan language family. In Dixon, R.M.W., Aikhenvald, A.Y. (eds.), Areal Diffusion and Genetic Inheritance: Case Studies in Language Change. Oxford: Oxford University Press, 225-254.
[22] Lee, Y.-C., Xu, Y. 2010. Phonetic realization of contrastive focus in Korean. Proceedings of Speech Prosody 2010, Chicago, 100033, 1-4.
[23] Liu, F., Xu, Y., 2005. Parallel encoding of focus and interrogative meaning in Mandarin intonation. Phonetica 62, 70-87.
[24] Mixdorff, H. 2004. Quantitative tone and intonation modeling across languages. Proceedings of International Symposium on Tonal Aspects of Languages: With Emphasis on Tone Languages Beijing, 137-142.
[25] O’Connor, J.D., Arnold, G.F. 1961. Intonation of Colloquial English. London: Longmans.
[26] Palmer, H.E. 1922. English Intonation, with Systematic Exercises. Cambridge: Heffer.