Possible Numeral Evolutions Prior to Proto-Sino-Tibetan
Quinary numerals in Pre-Proto-Sino-Tibetan
The high resemblance between Proto-Sino-Tibetan (PST) numeral *g/s-ni-s “two” and *s-ni-s “seven” has led to a deduction that Pre-Proto-Sino-Tibetan (Pre-PST) was probably retaining a quinary numeral system (since 7 = 5 + 2), where “6” “7” “8” “9” were derived from forms like “5 + 1” “5 + 2” “5 + 3” “5 + 4”. However, it is not apparent to recover any form of “five” from *s-ni-s “seven”, so do *d-k-ruk “six”, *b-r-gjat ~ b-g-rjat “eight” and *d/s-kəw “nine” (as to represent /j/, STEDT convention ⟨y⟩ is converted into IPA convention ⟨j⟩).[1] But, by taking quinary as an assumption and PST numerals as contractions from Pre-PST disyllabic forms, a possible evolutionary scheme could be obtained.
Listing major PST numerals reconstructed in STEDT, there are:[2]
“1”: *tjak ~ g-t(j)ik “6”: *d-k-ruk *kat *ʔit ~ jat “2”: *g/s-ni-s “7”: *s-ni-s “3”: *g-sum “8”: *b-r-gjat ~ b-g-rjat “4”: *b-ləj “9”: *d/s-kəw “5”: *l/b-ŋa “10”: *ts(j)i(j) ~ tsjaj *s-ra *gip
Superficially, besides the very similar *g/s-ni-s “two” and *s-ni-s “seven”, certain similarities can be found between “five”, small numbers and their “+5” counterparts, as well as in between those large numbers. For instance, the most common “one” *tjak ~ g-t(j)ik shares *-k with *d-k-ruk “six”, and *d-k-ruk “six”, *b-r-gjat ~ b-g-rjat “eight”, *d/s-kəw “nine” all contain *k-/*g- which may be related to *l/b-ŋa “five”. Since “ten” has multiple diverse forms which would be messy to deal with, this article mainly focuses on these numerals:
“5”: *l/b-ŋa “1”: *tjak ~ g-t(j)ik “6”: *d-k-ruk “2”: *g/s-ni-s “7”: *s-ni-s “3”: *g-sum “8”: *b-r-gjat ~ b-g-rjat “4”: *b-ləj “9”: *d/s-kəw
Notice *r- in *d-k-ruk “six” and *b-r-gjat ~ b-g-rjat “eight”, with quinary assumption they are expected to be related to *t- (~ *d-) in *tjak ~ g-t(j)ik “one” and *s- (~ *z-) in *g-sum “three” somehow. Similar correlations can indeed be found in PST (listing with Old Chinese (OC) reflexes reconstructed by Baxter-Sagart (2014)[3,4]):
PST *s-tjaŋ “upper part” “rise” “raise” “top” “summit”[5] → OC 上 /*Cə-daŋʔ/ “ascend”, 上 /*m-daŋʔ/ “raise” “deem superior”, 上 /*daŋʔ-s/ “top” “supreme”, 登 /*tˤəŋ/ “ascend” “step up”, 等 /*tˤəŋʔ/ “step (of stairs)” “classify”, 烝 /*təŋ/ “offer (gift, sacrifice)”, 承 /*m-təŋ/ “receive”, 升 /*s-təŋ/ “rise (v.)”, 頂 /*tˤeŋʔ/ “top” cf. PST *(z)rjaŋ “uncle” “superior”[6] → OC 尚 /*daŋ-s/ “noble”, 丈 /*draŋʔ/ “length unit (= 10 feet)” “older male”
Where *s-t- and *z-r- were probably derived from an earlier common ancestor, and *(z)rjaŋ seems to be more “contracted” than *s-tjaŋ . If the relation between alveolar stops/fricatives and *r holds, *-rVk in “six” could be allofamically related to *-tVk in “one”, and *-rVt in “eight” could be allofamically related to *-sVm in “three”. As for the coda difference between “three” and “eight”, *-m ~ *-t allofam can be found in PST as well — they were probably both derived from an earlier *-p (listing with OC and Proto-Lolo-Burmese (PLB) reflexes):
PST *s-nap “enter”, *s-ni(ː)p ~ r/s-njap ~ s-nu(ː)p “pinch” “squeeze” “press” “oppress” “sink into” “submerge” “enter”,*s-nem ~ s-njam “low” “soft”[7] → OC 入 /*nup/ “enter”, 內 / 納 /*nˤup/ “bring in”, 內 /*nˤup-s/ “inside”, 躡 /*nrep/ “trample”, 攝 /*kə.n̥ep/ “catch”, 溺 /*nˤewk/ “sink in water”, 西 /*s-nˤər/ “west” → PLB *ʔnip “squeeze” “press” cf. PST *s-njit “squeeze”[8] → PLB *ʔnjit “squeeze” “press”
Then, the actual form of Pre-PST “five” is to be discussed. Considering *ŋ- in PST “five”, *k- in PST “six” “nine” and *g- in PST “eight” are all expected to be derived from it, Pre-PST “five” might have had a velar stop or nasal initial. Besides, there do not seem to be evidence on any definite coda for “five”, so this article treats Pre-PST “five” as a simple open syllable *GV , where *G denotes the possible velar initial that derived *ŋ-/*g-/*k- in PST.
Since *G- became *ŋ- in *l/b-ŋa “five”, a contraction from *GV-nV might have undergone similar changes. That is, *GV-n- might contract into *G-n- , then became *ŋ-n- which would be highly possible to assimilate into a single nasal. This could explain why there are significant similarities between PST “two” *g/s-ni-s and “seven” *s-ni-s with no apparent traces of “five” — the “five” might have become a nasal prefix, then completely assimilated into nucleus *-ni- .
Thus, a speculative phylogenetic history for PST numerals could be proposed (notice the denoted consonants prior to PST do not necessarily represent their face values — some of which might have been changing drastically, but for most cases there is no recognizable evidence):
Pre-PST PST “1”: *-tVk- → *-tVk- → *-tVk → *tjak ~ g-t(j)ik “6”: *-GV-dVk- → *-G-dVk- → *-k-rVk → *d-k-ruk
“2”: *-nV- → *-nV- → *-nV- → *g/s-ni-s “7”: *-GV-nV- → *-G-nV- → *-ŋ-nV- → *s-ni-s
“3”: *-sVp- → *-sVp- → *-sVm → *g-sum “8”: *-GV-zVp- → *-G-zVp- → *-g-rVt → *b-r-gjat ~ b-g-rjat
“4”: *-lV- → *-lV- → *-lVj → *b-ləj “9”: *-GV-lV- → *-G-lV- → *-k-lVw → *d/s-kəw
“5”: *-GV → *-GV → *-ŋV → *l/b-ŋa
There are still affixes in PST numerals remain unexplained, say *g- in “one” “two” “three”, *s- in “two” “seven” “nine”, *b- in “four” “five” “eight”, *l- in “five”, *d- in “six” “nine”, and *-s in “two” “seven”, which may have implied more complex syllabic structures and evolutionary histories. Besides that, notice the exact parts in PST reflexes — denoted in bold — that were directly inherited from their hypothetical Pre-PST ancestors are all of continuous segments, and the consonantal changes involved are generally homorganic or near-homorganic. This could be a good sign of plausibility, though alternative possibilities exist as well.
Sinitic, Kuki-Chin and Pre-PST monosyllabic roots
By now, pharyngeal theory has usually been considered most plausible on explaining Old Chinese type-A/B distinction, i.e. 非三等 (fēi sānděng) “non-division III” vs. 三等 (sānděng) “division III” distinction, by which at least a majority of type-A syllables were characterized by pharyngealized initials that led them to become generally more “open” than unpharyngealized type-B syllables during Hàn period (汉代). In order to explain the origin of pharyngealization along with the correlations between type-A syllables and Kuki-Chin long vowels, Sagart & Baxter (2016)[9] proposed a hypothesis attributing to characteristics dating back to Pre-PST period. Briefly speaking, Pre-PST might have prevented the existence of monomoraic free words like in Austronesian and Austroasiatic, therefore monosyllabic words were realized as geminate vowels so as to avoid being monomoraic. By evolving into PST, those monosyllables became *CVʕVC and other syllables became unmarked *CVC , then afterward evolved into long *CVVC vs. short *CVC in Proto-Kuki-Chin (PKC) and pharyngealized *CˤVC vs. unpharyngealized *CVC in Old Chinese. So, now this article is to apply this hypothesis to PKC and OC numerals[2-4] in order to speculate on possible syllabic characteristics of their Pre-PST roots:
PST PKC reflex long vowel? OC reflex type-A? “1” *tjak ~ g-t(j)ik — — 隻 /*tek/ No “1” *kat *khat ~ ʔat ~ hat No — — “1” *ʔit ~ jat — — 一 /*ʔit/ No “2” *g/s-ni-s *niʔ ~ hniʔ No 二 /*nij-s/ No “3” *g-sum *thum No 三 /*s.rum/[a] No?[a] “4” *b-ləj *lii Yes?[b] 四 /*s.lij-s/ No “5” *l/b-ŋa *ŋaa Yes 五 /*C.ŋˤaʔ/ Yes “6” *d-k-ruk *ruk No 六 /*k.ruk/ No “7” *s-ni-s *sa-riʔ No 七 /*t͡sʰit/ No “8” *b-r-gjat ~ b-g-rjat *riat No?[b] 八 /*pˤret/ Yes “9” *d/s-kəw *kua No?[b] 九 /*kuʔ/ No “10” *ts(j)i(j) ~ tsjaj — — — — “10” *s-ra *hraa Yes — — “10” *gip — — 十 /*t.ɡəp/ No “10” — *soom Yes — —
[a] It is possible that OC 三 “three” has had varied pronunciations, since Middle Chinese (MC) 三 pronounced /sɑm/ — which seems to be quite “irregular” comparing to OC rhyme. By MC pronunciation, an OC /*sˤam/ is expected. [b] Vocalizations of PST *j *w into PKC *i *u probably have had “overriding” effects on PKC vowel length, leaving original vocalic characteristics being indetermined.
It could be observed that most numerals, including those small ones are dominantly “unmarked” (i.e. with short PKC vowel and type-B OC reflexes), indicating multisyllabic Pre-PST roots. And, the only numeral that definitely evolved into both long PKC vowel and type-A OC reflexes is *l/b-ŋa “five” — indicating a monosyllabic Pre-PST root which is consistent with the hypothesis made by this article above.
The simplicity of “five” may imply that it might have arisen from certain basic concepts related to “five”, say, “fingers” “toes” “hand” “foot” “palm” “sole” etc. — under which there are considerably abundant forms reconstructed for PST.[10] Additionally, the nuclei of PST “two” *g/s-ni-s and “four” *b-ləj are phonetically similar, which may be a trace of even earlier phylogenesis — specifically, when there were only “one” “two” and “three”, it was at least a logical possibility to derive a “four” from “two” by some means since either 2 + 2 or 2 × 2 equals 4.
Basic numerals in Mainland Southeast Asia languages
Small numerals, especially “two” “three” “four” “five” are among the most stable vocabularies during language evolution. They usually feature exceptional life expectancies longer than the reconstructed histories of nowaday language families — and even possibly longer than the actual histories of those numerals themselves.[11] Therefore, this article is now interested in whether PST/Pre-PST basic numerals have even prior traces, along with those in neighboring related language families.
Summarizing native numerals (i.e. could not be definitely attributed to later borrowing) in major language families of Mainland Southeast Asia (MSEA) linguistic area, there are:
Proto-Sino-Tibetan (PST)
Proto-Hmong-Mien (PHM) “2” “3” (“1” and “4” - “10” were from Sino-Tibetan)[12]
Proto-Austroasiatic (PAA)[13]
Proto-Austronesian (PAN)[14]
Proto-Kra-Dai (PKD) ├… │ ├Proto-Kra (PKr)[15] │ └… └… ├Proto-Hlai (PHl)[16] └…
Then listing numerals from “one” to “five”, there are:
PST PHM PAA PAN PKD PKr PHl “1” *tjak ~ g-t(j)ik *muuy ~ muəy ~ muuɲ *esa *tʂəm C *tɕʰɯ: *kat *isa *ʔit ~ jat “2” *g/s-ni-s *ʔu̯i *ɓaar *duSa *sa A *hlu:ʔ “3” *g-sum *pjɔu *piʔ *telu *tu A *tʃʰwuʔ *ʔu(u)y ~ ʔuəy *ʔ[ə]y “4” *b-ləj *punʔ ~ puən[ ] ~ pan[ ] *Sepat *pə A *tʃʰa:wʔ “5” *l/b-ŋa *p[ɗ]am *lima *r-ma A *hma: *suun ~ suən ~ sən
It could be recognized that PKr and PHl basic numerals are aligned with PAN ones, which is consistent with the close relations between Kra-Dai and Austronesian. Aside from this, there do not seem to be apparent systematic connections, though a few scattered similarities could be found. Actually, this is an expected result since numeral system was a quite late invention throughout the history of anatomically modern humans, as only after humans had entered agricultural societies did numerals become of importance — which did not happen until the beginning of Neolithic. If there were not significant phonological changes resulting in cognate numerals becoming drastically distinct, it is logical to presume that the basic numerals of Sino-Tibetan, Hmong-Mien, Austroasiatic and Austronesian/Kra-Dai might have emerged independently in history.
可能的早于原始汉藏语的数词演化过程
前原始汉藏语中的五进制数词系统
原始汉藏语数词 *g/s-ni-s “二” 与 *s-ni-s “七” 的高度相似使得语言学家认为前原始汉藏语可能曾经有一个五进制数词系统（考虑 7 = 5 + 2）——也就是说，“6” “7” “8” “9” 分别是从类似于 “5 + 1” “5 + 2” “5 + 3” “5 + 4” 的形式派生出的。然而，从 *s-ni-s “七” 中看上去并不能还原出任何表示 “五” 的成分，*d-k-ruk “六”、*b-r-gjat ~ b-g-rjat “八” 及 *d/s-kəw “九” 亦是如此（在表示 /j/ 时本文使用国际音标标准 ⟨j⟩ 而非STEDT标准 ⟨y⟩）；[1] 不过，通过直接假设原始汉藏语的数词是由更早的五进制双音节形式缩约而来的，我们还是能推测出一个可能的数词演化图景。
STEDT列出的主要的原始汉藏语数词包括：[2]
“1”: *tjak ~ g-t(j)ik “6”: *d-k-ruk *kat *ʔit ~ jat “2”: *g/s-ni-s “7”: *s-ni-s “3”: *g-sum “8”: *b-r-gjat ~ b-g-rjat “4”: *b-ləj “9”: *d/s-kəw “5”: *l/b-ŋa “10”: *ts(j)i(j) ~ tsjaj *s-ra *gip
表面上，除去非常相似的 *g/s-ni-s “二” 与 *s-ni-s “七”，可以发现 “五”、小数词同其对应的 “+5” 大数词之间，还有在大数词的内部均存在一些相似之处。例如，最常见的 “一” *tjak ~ g-t(j)ik 含有与 *d-k-ruk “六” 相同的 *-k ， *d-k-ruk “六”、 *b-r-gjat ~ b-g-rjat “八”、 *d/s-kəw “九” 均含有可能同 *l/b-ŋa “五” 有关的 *k-/*g- 。考虑 “十” 拥有难以处理的多个不同形式，本文主要关注以下数词之间的联系：
“5”: *l/b-ŋa “1”: *tjak ~ g-t(j)ik “6”: *d-k-ruk “2”: *g/s-ni-s “7”: *s-ni-s “3”: *g-sum “8”: *b-r-gjat ~ b-g-rjat “4”: *b-ləj “9”: *d/s-kəw
注意 *d-k-ruk “六” 与 *b-r-gjat ~ b-g-rjat “八” 中的 *r- ——依五进制数词假设其应该同 *tjak ~ g-t(j)ik “一” 中的 *t- (~ *d-) ，及 *g-sum “三” 中的 *s- (~ *z-) 有关。实际上，在原始汉藏语中能发现相似的联系（附可能派生出的上古汉语词汇，白一平-沙加尔 (2014) 构拟[3,4]）：
原始汉藏语 *s-tjaŋ “上部” “上升” “提升” “顶部” “山顶”[5] → 上古汉语 上 /*Cə-daŋʔ/ “上升”, 上 /*m-daŋʔ/ “提升”, 上 /*daŋʔ-s/ “顶部” “上级”, 登 /*tˤəŋ/ “登上”, 等 /*tˤəŋʔ/ “台阶的级” “等级”, 烝 /*təŋ/ “进献”, 承 /*m-təŋ/ “接受”, 升 /*s-təŋ/ “上升”, 頂 /*tˤeŋʔ/ “顶部” cf. 原始汉藏语 *(z)rjaŋ “叔叔” “上级”[6] → 上古汉语 尚 /*daŋ-s/ “高尚”, 丈 /*draŋʔ/ “长度单位（= 10尺）” “老年男性”
其中 *s-t- 和 *z-r- 可能派生自更早的共同祖先，且 *(z)rjaŋ 似乎要比 *s-tjaŋ 更加 “缩约”。如果认同 *r 与齿龈塞音/擦音的联系，那么 “六” 中的 *-rVk 就能同 “一” 中的 *-tVk 联结，“八” 中的 *-rVt 亦能同 “三” 中的 *-sVm 联结。至于 “三” 和 “八” 在音节尾上的差异， *-m ~ *-t 也出现在原始汉藏语中——二者可能共同派生自更早的 *-p（附可能派生出的上古汉语和原始缅彝语词汇）：
原始汉藏语 *s-nap “进入”, *s-ni(ː)p ~ r/s-njap ~ s-nu(ː)p “挤压” “按压” “压迫” “沉入” “淹没” “进入”,*s-nem ~ s-njam “低的” “软的”[7] → 上古汉语 入 /*nup/ “进入”, 內 / 納 /*nˤup/ “放入”, 內 /*nˤup-s/ “内部”, 躡 /*nrep/ “踩”, 攝 /*kə.n̥ep/ “拿取”, 溺 /*nˤewk/ “淹没”, 西 /*s-nˤər/ “西方” → 原始缅彝语 *ʔnip “挤压” “按压” cf. 原始汉藏语 *s-njit “挤压”[8] → 原始缅彝语 *ʔnjit “挤压” “按压”
接下来讨论的是前原始汉藏语 “五” 的形式。考虑其可能派生出了原始汉藏语 “五” 中的 *ŋ- 、“六” “九” 中的 *k- 和 “八” 中的 *g- ，前原始汉藏语 “五” 可能存在一个软腭塞音或软腭鼻音的音节首。此外，似乎没有证据表明其存在任何明确的音节尾，故本文将 “五” 视为一个简单的开音节 *GV ，其中 *G 代表派生出了原始汉藏语 *ŋ-/*g-/*k- 的可能的软腭塞音或鼻音。
考虑 *G- 变为了 *l/b-ŋa “五” 中的 *ŋ- ， *GV-nV 也可能产生相似的音变。也就是说， *GV-n- 可能先缩约成 *G-n- 后变为 *ŋ-n- ，然后很可能会同化成一个单一的鼻音首。这可以解释为什么原始汉藏语中的 “二” *g/s-ni-s 与 “七” *s-ni-s 在非常相似的同时缺少 “五” 的痕迹—— “七” 中所包含的 “五” 可能先演变为了一个鼻音前缀，然后完全同化进了音节核 *-ni- 。
由此可以提出一个推测性的原始汉藏语数词的发生历史（注意原始汉藏语以前的辅音符号可能不代表其真实音值——在漫长的演变历史中一些辅音可能经历过巨大变化，但在绝大多数情况下都没有留下可分辨的证据）：
前原始汉藏语 原始汉藏语 “1”: *-tVk- → *-tVk- → *-tVk → *tjak ~ g-t(j)ik “6”: *-GV-dVk- → *-G-dVk- → *-k-rVk → *d-k-ruk
“2”: *-nV- → *-nV- → *-nV- → *g/s-ni-s “7”: *-GV-nV- → *-G-nV- → *-ŋ-nV- → *s-ni-s
“3”: *-sVp- → *-sVp- → *-sVm → *g-sum “8”: *-GV-zVp- → *-G-zVp- → *-g-rVt → *b-r-gjat ~ b-g-rjat
“4”: *-lV- → *-lV- → *-lVj → *b-ləj “9”: *-GV-lV- → *-G-lV- → *-k-lVw → *d/s-kəw
“5”: *-GV → *-GV → *-ŋV → *l/b-ŋa
仍有一些原始汉藏语词缀没有解释，如 “一” “二” “三” 中的 *g- 、“二” “七” “九” 中的 *s- 、“四” “五” “八” 中的 *b- 、“五” 中的 *l- 、“六” “九” 中的 *d- 、“二” “七” 中的 *-s ；它们可能意味着曾经存在过的更加复杂的音节结构和演化历史。除了这点，注意上面列出的各个原始汉藏语形式中直接继承自其假设的前原始汉藏语词根的部分（用粗体标示）都是连续的区段，而列出的辅音演变都是同部位的或接近同部位的音变——这可能代表着一定的合理性，尽管其他可能性同样存在。
汉语支、库基-钦语支和前原始汉藏语单音节词汇
目前，咽音理论一般被认为是解释汉语中三等字—非三等字分化的最优选择。在咽音理论中，至少绝大部分的非三等字在上古时期都拥有咽化的音节首，使其在汉代时往往演变得比非咽化的三等字要更加 “开口化”。为了解释咽化特征的来源及其同库基-钦语支长元音的相关性，沙加尔 & 白一平 (2016)[9] 提出了一种追溯至前原始汉藏语的假设。简略来说，类似于在南岛语系和南亚语系中出现的现象，前原始汉藏语可能会在语流中规避单音拍自由词的存在，于是本来的单音节词就会以双胞元音的方式来实现从而避免其成为单音拍。随着前原始汉藏语演变为原始汉藏语，这些单音节词变为了 *CVʕVC 而其他词汇则变为了无标记的 *CVC ，并在随后演变成了原始库基-钦语中的长元音 *CVVC 与短元音 *CVC 的对立以及上古汉语中的咽化 *CˤVC 与非咽化 *CVC 的对立。所以，现在本文将基于这个假设来考察原始库基-钦语数词与上古汉语数词[2-4]以推测其前原始汉藏语词根的可能特征：
原始汉藏语 原始库基-钦语 长元音？ 上古汉语 非三等？ “1” *tjak ~ g-t(j)ik — — 隻 /*tek/ 否 “1” *kat *khat ~ ʔat ~ hat 否 — — “1” *ʔit ~ jat — — 一 /*ʔit/ 否 “2” *g/s-ni-s *niʔ ~ hniʔ 否 二 /*nij-s/ 否 “3” *g-sum *thum 否 三 /*s.rum/[a] 否？[a] “4” *b-ləj *lii 是？[b] 四 /*s.lij-s/ 否 “5” *l/b-ŋa *ŋaa 是 五 /*C.ŋˤaʔ/ 是 “6” *d-k-ruk *ruk 否 六 /*k.ruk/ 否 “7” *s-ni-s *sa-riʔ 否 七 /*t͡sʰit/ 否 “8” *b-r-gjat ~ b-g-rjat *riat 否？[b] 八 /*pˤret/ 是 “9” *d/s-kəw *kua 否？[b] 九 /*kuʔ/ 否 “10” *ts(j)i(j) ~ tsjaj — — — — “10” *s-ra *hraa 是 — — “10” *gip — — 十 /*t.ɡəp/ 否 “10” — *soom 是 — —
[a] 上古汉语 三 有可能存在不同读音，考虑中古汉语 三 的读音为 /sɑm/ ——对于上古韵部而言是个非常 “不规则” 的结果。单纯从中古音逆推得出的上古音应为 /*sˤam/。 [b] 原始汉藏语 *j *w 元音化成原始库基-钦语 *i *u 的过程可能 “覆盖” 了本来的元音长度，使得原本的元音长度特征变得无法判断。
可以发现包括小数字在内的大部分数词都是以 “无标记”（即派生出原始库基-钦语短元音与上古汉语三等字）为主的，表明了多音节的前原始汉藏语词根；而，唯一一个明确派生出了原始库基-钦语长元音与上古汉语非三等字的数词就是 *l/b-ŋa “五” ——表明了一个单音节的前原始汉藏语词根，符合本文在上一部分作出的假设。
数词 “五” 的简单性也许意味着其可能是从与 “五” 相关的某个基本概念衍生出来的，比如 “手指” “脚趾” “手” “脚” “手掌” “脚掌” 等——在这些概念下存在为数众多构拟出的原始汉藏语词根。[10] 另外，原始汉藏语 “二” *g/s-ni-s 与 “四” *b-ləj 的音节核也比较相似，也许意味着某种更加远古的发生过程；详细来说，当只存在数词 “一” “二” “三” 时，以某种方式从 “二” 导出 “四” 至少是一种逻辑上的可能性，考虑无论是 2 + 2 还是 2 × 2 均等于 4 。
大陆东南亚语言中的基本数词
小数词，尤其是 “二” “三” “四” “五” 是语言演化中最为稳定的词汇之一；它们往往拥有极长的预期寿命——长于现今各个语系的构拟历史，甚至可能长于这些数词本身的历史。[11] 因此，本文现在打算考察在邻近的关联语系中是否存在同原始汉藏语/前原始汉藏语基本数词有关的痕迹。
总结大陆东南亚语言联盟中的原生数词（即不被认为是后来借入的数词），有：
原始汉藏语
原始苗瑶语 “2” “3”（“1” 和 “4” - “10” 借自汉藏语系）[12]
原始南亚语[13]
原始南岛语[14]
原始侗台语 ├… │ ├原始仡央语[15] │ └… └… ├原始黎语[16] └…
然后列出从 “一” 至 “五” 的数词，有：
原始汉藏语 原始苗瑶语 原始南亚语 原始南岛语 原始侗台语 原始仡央语 原始黎语 “1” *tjak ~ g-t(j)ik *muuy ~ muəy ~ muuɲ *esa *tʂəm C *tɕʰɯ: *kat *isa *ʔit ~ jat “2” *g/s-ni-s *ʔu̯i *ɓaar *duSa *sa A *hlu:ʔ “3” *g-sum *pjɔu *piʔ *telu *tu A *tʃʰwuʔ *ʔu(u)y ~ ʔuəy *ʔ[ə]y “4” *b-ləj *punʔ ~ puən[ ] ~ pan[ ] *Sepat *pə A *tʃʰa:wʔ “5” *l/b-ŋa *p[ɗ]am *lima *r-ma A *hma: *suun ~ suən ~ sən
可以发现原始仡央语、原始黎语的基本数词同原始南岛语对应，符合侗台语系与南岛语系的紧密联系；除此之外的数词似乎没有明显的系统性对应关系，尽管可以找到少许分散的相似性。实际上，这是一个期望中的结果，考虑数词系统在整个晚期智人史中是一个相当 “晚近” 的发明：只有当新石器时代开始、人类进入农业社会之后数词系统才变得重要。如果不是因为剧烈的语音变化使得同源数词变得完全不同，逻辑上可以推测汉藏语系、苗瑶语系、南亚语系、南岛语系/侗台语系的基本数词在历史上可能是各自独立产生的。
This work is licensed under a Creative Commons Attribution 4.0 International License. 本作品采用知识共享署名 4.0 国际许可协议进行许可。
[1] STEDT, “#2505 PTB *s-ni-s SEVEN” [2] STEDT, “9.9 Numbers and numeric expression” [3] W. H. Baxter (白一平) & L. Sagart (沙加尔) (2014), Old Chinese: A New Reconstruction, ISBN: 9780199945375 [4] W. H. Baxter (白一平) & L. Sagart (沙加尔) (2014), “The Baxter-Sagart reconstruction of Old Chinese” [5] STEDT, “#2725 PTB *s-tyaŋ UPPER PART / RISE / RAISE / TOP / SUMMIT” [6] STEDT, “#2640 PTB *(z)ryaŋ UNCLE / SUPERIOR” [7] STEDT, “#2517 PTB *s-n(i/u)(ː)p/m ⪤ *r/s-nyap/m PINCH / SQUEEZE / PRESS / OPPRESS / SUBMERGE / SINK INTO / WEST / LOW / SOFT” [8] STEDT, “#5525 PTB *s-nyit SQUEEZE” [9] L. Sagart (沙加尔) & W. H. Baxter (白一平) (2016), “A Hypothesis on the Origin of Old Chinese Pharyngealization (上古汉语咽化声母来源的一个假设)”, Bull. Chin. Linguist. 9(2):179-189, DOI: 10.1163/2405478X-00902002 [10] STEDT, “1.5 Limbs, Joints, and Body Measures” [11] M. Pagel & A. Meade (2017), “The deep history of the number words”, Philos. Trans. Royal Soc. B 373:20160517, DOI: 10.1098/rstb.2016.0517 [12] M. S. Ratliff (2010), Hmong-Mien Language History, ISBN: 9780858836150, ABVD, “Proto-Hmong-Mien” [13] H. Shorto, et al. (2006), A Mon-Khmer Comparative Dictionary, ISBN: 9780858835702, ABVD, “Proto-Mon-Khmer” [14] R. A. Blust (1999), “Subgrouping, Circularity and Extinction: Some Issues in Austronesian Comparative Linguistics”, ABVD, “Proto-Austronesian” [15] W. Ostapirat (2000), “Proto-Kra”, Linguist. Tibeto-Burman Area 23(1):1-251, ABVD, “Proto-Kra” [16] P. K. Norquest (2007), “A Phonological Reconstruction of Proto-Hlai”, ABVD, “Proto-Hlai (Norquest)”
