IV. BIOLOGICAL FOUNDATIONS OF HISTORY AND SITRIBUTION OF NATURAL LANGUAGES
(1) Theoretical Foundations
Most of the general phenomena of natural languages, especially the reasons for change and relative constancy follow logically from the theory postulated here, particularly form propositions (5), (12) and (13), that is, the concepts of latent and realized structure, of resonance, and of individual variations.
在自然語言中最普遍的現象，就是產生共鳴的潛性結構(latent structure)和個體變異的顯性結構(realized structure)，本書試圖以假設性的邏輯理論加以解釋，尤其是對變異以及相對恆定(relative constancy) 的因素，所以特別提出了第(5)、(12)、(13)項的論點。
The types of changes encountered in the recorded history of languages are too well-know to require further comment here. The relative constancies are also fairly obvious though they are more frequently ignored in discussions of language. Features that remain constant throughout recorded language history are identical with language universals encountered in modern languages. Thus, we find changing inventories of phonemes but complete constancy in the basic phonematic mode of signaling. (All languages are and have been phoneme-language throughout the documented past. Sound analogy or sound imitation is never encountered un-phonematized.) Although lexica, from-class assignment, and syntactically marked groupings of words may differ through history and place, all languages are concerned with essentially similar aspects of the environment. (There is no language in which we cannot give directions, tell about a past event, describe other persons’ behavior, etc.).
在歷史的紀錄裡，語言變化的類型已經眾所周知，所以不必在此多加贅述，而語言的相對恆定雖然較容易在研究語言時被忽略，但目前也已有一番觀察。在歷史的紀錄中，恆定的特徵在所有語言中是普遍相同的，現代語言也不例外，因此我們發現音素的存留雖有改變，但是在發出信號的基本音位調式上是完全恆定的，(由歷史紀錄可知，所有語言都是以音素為本的語言，聲音的類比和聲音的模仿永遠都是音素化的。) 雖然詞性分類的字彙和句法規範的詞句，可能因為歷史與地方而有所不同，但所有的語言都和環境的本質相似。 (所有的語言我們都能用來指示方向、述說過去發生的事、描述別人的行為等等。)
The syntax of all languages is of the same basic type: utterances consist of concatenated morphemes and the concatenation is never random. Words and morphemes are always assignable to functional categories; sentences in all languages may be judged as grammatically acceptable or unacceptable by criteria of underlying structure. We see that the superficial structure of languages undergoes noticeable changes in history and thus results in diversification of languages over the face of the earth, whereas there are powerful indications that the underlying structure preserves its essential type over very long periods of time and among all speech communities. Our theory attempts to relate these two levels of language structure to what we have called latent and realized structure, terms that refer to the structure of behavior instead of the structure of the behavioral product.
所有語言的句法都有相同的基本類型，語句(utterance)裡包含一連串的語素(morpheme)，但這種串連並不是任意編排的。詞和語素永遠都能依其詞性分門別類；利用句法深層結構(underlying structure)的標準，所有語言的句子都能判定是否合乎文法規則。我們發現語言中的表層結構(superficial structure) 在歷史洪流中會有所改變，進而造成全球語言的多樣性，不過深層結構這個強而有力的規範，長久保存了所有的語言群落的結構本質。本書的理論就是要試著連結潛性結構)和顯性結構的關係，也就是語言運轉狀態的結構而非由語言行為產出的結構。
(a) Source, inhibitors, and determinants of change. The system of biological replication does not work with ideal precision; random variations occur which , over the long run, would have a totally self-cancelling and , therefore, leveling and difference-eroding effect unless some factors of bias were operating selectively upon this bounty of variations. This is true of life, in general, and of all aspects of organisms. Thus the source for change is in the self-replication mechanism itself, and our problem therefore is not why a given characteristic of the species (in our case aspects of communication behavior) changes but what inhibits the change.
(a) 改變的來源、限制、決定因素：生物的複製系統並不是百分之百精確的，隨機發生的變異，最後都會全部自我銷毀，除非這些混亂的因素能被選擇性的運作與接受，這些擊毀和磨損語言的變異才具有影響力，普遍來說，這就是真實的生命，就是生物體的所有面貌。改變的來源就是自我複製的過程，但我們的問題不在於為什麼種族本有的特徵會改變(這裡指的是溝通行為)，而是到底是甚麼限制了改變。
There must be at least two types of inhibitors: an extremely powerful one that holds the capacity for latent structure relatively constant over a time span of probably no less than 50,000 years, and a very weak one that permits changes in outer form and realized structure to occur within just a few generations. Given the source of variation and inhibiting factors that permit either only very slow or quite rapid changes, what might the nature of the factors be that determines the direction of change? Since there are two types of inhibitors, it is reasonable to assume that the direction-giving determinants are also of two different kinds. The slow changes in the capacity for latent structure are undoubtedly given direction by those factors that are generally held to be responsible for directionality in evolution (see Chapter Six). They combine to build up selection pressure that is either directly or indirectly responsible for the shaping of the species. The second type of determinants are probably a combination of social factors (cultural invasion, diffusion, and prestige factors), internal restructuring of the language pattern as a whole (for instance, the substitution of word order for inflection, the introduction of new words due to loss of semantic specificity of older words, or the development of new phonemic contrasts due to fusion of older contrasts), and perhaps uncontrolled exercise of freedom for random variations. In the following pages we shall not be concerned with the factors that determine direction of change but primarily with the nature of the inhibitors.
至少有兩種限制：第一種限制是極具力量並控制潛性結構的功能，這種限制相當穩定的留存約有五萬年之久；另一種限制則較為薄弱，僅發生在這幾個世代，其中外在形式與淺層結構的改變都是可以被接受的。造成變異來源與限制的因素，可以是很緩慢或是極快速的，但到底是甚麼自然因素控制了決定改變的方向？既然有兩個種類的限制，想必決定性的控制因素也有兩種不同。潛性結構功能的緩慢改變，無疑的是受到負責引導演化的因素影響 (請參照第六章)，演化是由選擇的壓力所造成，直接或間接將各個種族塑型。第二種決定因素可能是由兩種因子所綜合而成，第一種是將兩者視為一個整體的社會因素(文化的入侵、傳播、地位等)和內在語言形式的重組(例如字詞替代的曲折變化；引入新詞取代失去特定語意的舊詞；融合舊有的音韻對比發展出新的音韻對比)，第二種則是不受控制的自由變異。接下來本書的焦點將從決定性的控制因素，轉移至自然的限制。
(b) Variance in capacity for latent and realized structure. Table 9.1 presents a rough classification of the most important types of possible variations. The table is self-explanatory. Alterations of latent structure may be due to cognitive deviations or due to variations of the language input, that is, breaking down of elements and resynthesizing them to form a realized structure, is fundamentally affected and thus the propensity for language or Chomsky’s competence is altered. On the other hand, in alterations in the capacity for actualization, competence remains close to the population mean, whereas aspects of performance are affected, primarily due to peripheral or generally incidental reasons. Peculiarities in the individual geometry of the vocal tract, or auditory peculiarities, or idiosyncrasies of general motor behavior, or other similar factors, will leave their imprint upon an individual’s language output, resulting in different styles of normal speech.
(b) 潛性和顯性結構功能的變異：表格9.1粗略整理了最重要的可能變異類型，這個圖表本身就有解釋的功能。潛性結構的改變可能是因為認知的偏離或是語言產出的變異，就是指分割成一個個要素之外，再重新合成產生顯性結構，因此對語言的歧異或Chomsky的語言能力論都有重要的影響。另一方面，在實際能力的改變之下，語言能力仍保持在一般的平均值，然而語言表現則受到影響，主要得歸因於周邊或偶發的理由，每個獨立個體奇特形狀的聲道、獨特的聽覺、一般運動行為的特有風格、或其他類似的因素，都會在他們的語言產出中留下印記，造成他們說話的不同風格。
(c) Tolerance for variance; the mechanism of all changes. A characterization of the variance is not sufficient to account for the evolutionary and historical behavior of languages. For this we must introduce the concept of tolerance. Consider deviations from the population mean in capacity for latent structure. If the deviations are marked, the individual is either incapable of building up language behavior and thus can neither understand nor speak, or it is at least theoretically possible that there is latent structure of an altered type so that the language that is built up is patterned upon a freak type of underlying structure, unique in its operation, resulting in the generation of types of rules that are incomprehensible to the rest of the community. In either case, this type of language cannot propagate through the biological community because there would be no individual or not enough individuals whose latent structure is similarly deformed to allow them to resonate efficiently to such deviant behavior. On the other hand, it might be postulated that the deviation is due to a genetically transmitted trait, so that deviant children could resonate to their deviant parents. In fact, language deficient families do exist, although it is not yet certain that the mechanism is necessarily the one postulated here. However, there is another reason why marked deviations of latent structure have a low chance for dissemination.
(c) 變異的接受；所有改變的機制：變異特性的描述並不足以說明語言的演化和過去行為，因此本書必須介紹「接受」(tolerance)的概念。試想偏離一般語言能力平均值的潛性結構，若這種偏離是顯著的，則這個個體不是無法建構語言行為，因而無法理解和說話，就是改變了潛性結構，然後在顯性結構上組合成反常的句式，而獨特的運作方式導致無法被其他群體理解的衍生法則，理論上是很有可能會發生的。任何一個例子當中，這種歧異的語言在生物群體裡將無法延續，因為沒有任何一個個體或是足夠的個體有這樣奇怪的潛性結構，讓他們和這樣歧異的行為產生有效的共鳴。另一方面，歧異或許能被假設是基因傳遞的特徵，所以有這種歧異的孩子能夠和有歧異的父母產生互動，雖然還無法完全確定這樣歧異機制的必然性，但事實上，有這種缺陷的家庭是存在的，然而還有另一個理由解釋為什麼潛性結構的歧異不太能傳播開來。
