ASPECTS OF LINGUISTICS

An utterance is conveniently viewed as operating on four levels: sound (phonology), form (morphology), syntax, and vocabulary (lexicon). No two utterances linguistically the same can be physically identical. That is, exact physical sameness cannot be achieved by the same speaker, let alone by different speakers, who have speech organs that naturally differ in shape and size, and who have their own peculiarities of enunciation. It follows that linguistic sameness of utterances cannot reside in physical sameness, but rather that speech events are regarded as “sames” if, despite their physical nonidentity, they convey the same message. This is so because any element on any one of the four levels is more than just an acoustic event. It represents a class of events, it attests an abstraction. Just as the human mind is capable of placing an infinite number of optical IMPRESSIONS into the class “hat,” having learned what characteristics define hat-ness, as it were, so also the infinitely numerous acoustic events hat are classified. The THREE sound-events, (incidentally represented by an equal number of letters) are recognized as attesting h-ness, a-ness, and t-ness. And in all utterances hats, the fourth elements, s, is recognized as a form-event attesting plural-ness. Linguistics deals with both the event (in Saussure’s terminology, parole) and the class to which the event belongs (langue), and it deals with them on any of the four levels of analysis, synchronically or diachronically.

语言学方面

我们可以方便地将话语视为在四个层次上的操作：声音（音韵学）、形式（形态学）、句法规则和词汇（词典）。在语言上相同的两个话语在物理上不可能完全相同。就是说，同一位演讲者无法完全实现相同的物理相同性，更不用说不同的演讲者，他们的言语器官在形状和大小上自然不同，并且有着自己的发音独特性。由此可见，话语的语言相同性无法存在于物理的相同性中，相反，如果他们传达了相同的信息，尽管他们在物理上没有相同性，那么讲演活动则被视为“相同”。之所以如此，是因为在四个层次上的任何一个级别的任何元素不仅仅只是一个声音事件。它代表一类事件。它证实了一种抽象。正如人类的心灵有能力将无限数量的视觉印象纳入类“帽子”中一样。在了解了定义帽子的特征后，可以说，帽子无限多的声学事件同样被归类了。三个声音事件（顺便用相同数量的字母表示）被公认为h-性质，a-性质和t-性质。而在帽子的所有话语中，第四个元素s, 被认可为证实复数性质的一种形式事件。语言学既涉及事件（用索绪尔的术语，言语）也涉及事件所属的（语言）类别，并在四个分析层次中的任何一个层面上处理它们，无论是共时的，还是历时的。

Synchronic Linguistics—Phonology. The sound-events are the subject of phonetics, and the sound-classes the subject of phonemics. (Transformational grammar does not acknowledge a separate phonemic level, preferring to state the utterance phonetically.)

共时语言学—音韵学。声音事件是语音学的主题，而声音类别是音位学的主题。（转换性语法不承认单独的音位级别，更喜欢根据发音来陈述话语。）

Phonetics deals with the sounds as events in physical terms, but several approaches are possible. Articulatory phonetics, the most commonly practiced and the oldest variety, states by means of what organs and by what employment of these organs—place and type of articulation—a given sound is brought forth. Acoustic phonetics describes the dislocation of air molecules, the “sound waves,” that an articulation produces and that impinge upon the eardrum. Whence they are transmitted to the nervous system and the brain. These sound waves can also be captured and rendered visible by recording devices, such as the oscilloscope and spectrograph. Perceptual phonetics identifies each sound by a set of so-called inherent, or distinctive, features that are listed in binary opposition, such as lax or tense, vocalic or non-vocalic, strident or mellow. These features, which complement rather than replace articulatory classifications, are discernible on the spectrogram as characteristic distributions of acoustic energy.

语音学将声音处理为物理术语事件，但不同的方法是可能的。发音语音学，最常用和最古老的种类，陈述通过什么器官和通过这些器官的用途—发音位置和类型—生产一个给定的声音。发音语音学描述了空气分子的转位，说话生产“声波”，并且对鼓膜生产影响。由此，它们被传送至神经系统和大脑。通过记录设备也可捕获这些声波，并使其可见，诸如示波器和光谱仪。感知语音学通过一套所谓固有的，或与众不同，在二元对立中列出的特征来识别每种声音，诸如松弛或紧张，含元音的或不含元音的，刺耳的或悦耳的。这些特征，补充而不是取代发音的分类，在声谱图上可辨别为声能的特征分布。

The normal orthography of most languages, especially that of English, may render the same sound by more than one symbol—for example, the k-sound of English in kind, corn, choir, pique, sack—or employ the same symbol for more than one sound—for example, the letter u in the English words but, butte, hurt, truth, pull. A phonetic alphabet was therefore designed using mostly symbols of the Latin alphabet, with some supplementary ones. Each symbol represents but one sound, and each sound has one written representation. It should be noted that even on this level of so-called narrow phonetic transcription each symbol stands for a class of phonetically very similar sounds; if one aimed at complete fidelity with the physical event no two sound-events could ever be rendered by the same symbol.

大多数语言的正字法，尤其是英语的正字法，可通过多个符号呈现相同的声音—例如，英语中kind, corn, choir, pique, sack的k-音—或者对多个声音使用相同的符号—例如，英语单词but, butte, hurt, truth, pull 中的字母u. 因此，音标的设计主要使用拉丁字母符号，包括一些补充的符号。每个符号只代表一个声音，而每个声音都有一个书写表示。应该注意的是，即使在这种所谓的狭义音标层级上，每个符号也代表一类语音学上非常相似的声音；如果一个声音完全忠实于物理事件，那么两个声音事件就不能用相同的符号来呈现。

Phonemics deals with the classes of sounds of a given language. The members of each class are functionally, linguistically equivalent, but physically different, though similar to one another. These subtypes occur either predictably as positional variants, or in free alternation. For an example of the first, an initial t in English, as in tone, is regularly pronounced with a puff of breath [th], whereas t after s, as in stone, remains mostly unaspirated. For an example of the second, the final t in English hatmay be aspirated [th], or non-aspirated [t], or not fully released [t]. These phonetically different subtypes of English t all convey the linguistic function of t-ness. Pronouncing tone with [t] may sound odd, but it does not alter the meaning of what has been said, and any of the THREE subtypes of t may be used in hat without even sounding odd. However, t-ness itself is opposed, in the English sound system, to d-ness and p-ness and all the other classes of speech sounds. Hence one says that the various phonetic subtypes of t belong to the same functional class and do not serve distinctive purposes. Only their replacement by another sound type, such as cone, bone, or hone for tone, produces different lexical items.

音位学涉及一种给定语言的声音类别。每个类别在功能上，在语言上都是等效的，但在物理上是不同的，虽然彼此相似。这些子类型要么是可预见的位置变体，要么是自由交替。第一个例子，如tone，英语中位于首位的t, 经常用一口气发音[th]，但在s之后的t，如stone，多半是不吸气。第二个例子，在英语hat中最后的可以吸气[th]，或不吸气[t]，或者不完全释放[t]。英语t的这些语音学上不同的子类型都传递了t-性质的语言功能。用t发音tone可能听起来奇怪，但它不改变所讲内容的含义，t的三种子类型中的任何一种都可用于hat，甚至听起来并不奇怪。然而，在英语语音系统中，t-性质本身与d-性质和p-性质以及所有其它的语音音素类别是相反的。因此人们说，t的各种语音子类型属于相同的功能类别，并且不用于特殊的目的。只有它们被其它声音类型所替换，诸如cone, bone, or hone for tone，才能产生不同的词汇项。

A class of functionally same sounds is called a phoneme, and the different subtypes belonging to that class are called allophones. (Most linguists operate with the phoneme in the same manner but its definitions diverge. The one implied here, seeing in the phoneme a class of speech sounds, goes back to the Leningrad linguist L. V. Scerba.) Accordingly, English has one phoneme /t/, with at least THREE allophones [t, th, t]. (Slant lines and brackets customarily indicate phonemes and allophones, respectively.) Indeed the speaker is normally not aware that he employs different allophones, but believes, significantly, that he utters the “same” sound. Of course, if a language employs [th] and [t] distinctively, if it has two words, called a minimal pair, [thi] and [ti], then it has two different phonemes /t/ and /th/. Each language employs, out of the numerous noises a person can make, a small number as phonemes, all of which occur according to combinational or phonotactic, rules. For example, no English word can begin with /ps/, despite the spelling psychology. All speech sounds are just noises, and the so-called speech organs all have primary biological functions. Fewer than one dozen phonemes per language would necessitate many very long words, and more than four dozen, though permitting short words, would overtax the human capacity for forming and perceiving sufficiently distinct sounds.

一类功能上相同的声音被称为音素，而属于该类的不同子类别被称为音位变体。（大多数语言学家以相同的方式处理音素，但其定义相异。这里隐含着，在音素中看到一类语音音素，可追溯到列宁格勒的语言学家L. V. 谢尔巴。）相应的，英语有一个音素/t/，至少有三个音位变体[t, th, t]，（斜线和括号习惯上分别表示音素和音位变体。）的确，说话人通常没有意识到他使用了不同的音位变体，但重要的是，相信他发出了“相同的”声音。当然，一种语言特殊使用了 [th] 和[t]，如果它有称为最小对的两个单词，那么它就有两个不同的音素/t/ 和 /th/。在一个人可以发出的众多响声中，每种语言都使用少量音素作为音素，所有这些音素都是根据组合或音位结构学，规则等发生的。例如，英语单词不能用/ps/开头，尽管有拼写心理学。所有的语音音素都只是噪音，而且所谓的发音器官都具有初级的生物学功能。每种语言只有不到一打的音素就需要许多很长的单词，而且超过四打，尽管允许用短词，但会过度消耗人类形成和充分感知不同声音的能力。

In addition to the phonemes, which occur successively in time and are therefore called segmental, there also exist nonsegmental (suprasegmental, prosodic) traits, which may or may not have a distinguishing (prosodemic) function. Unlike the distinctive features, they are not inherent but additive; hence their presence can be detected only in contrast to their absence in a given utterance. Thus a syllable is perceived as, say, accented only in comparison with surrounding nonaccented syllables. In other words, nonsegmental features exist in syntagmatic contrast, not in paradigmatic opposition. There are several such features: (1) Accent is the prominence given to a syllable through either loudness or pitch differentiation, frequently both. (2) Intonation distinguished, for example, It is raining from It is raining? (3) Quantity distinguishes short from long phonemes (in the vowels and consonants of classical Latin, for example). (4) Tone, by means of different pitch contours upon segmentally identical syllables, distinguishes words in so-called tone languages, such as Chinese and many African idioms.

除了音素以外，它们在时间上会连续出现，因此称为音段的，在那个点上也存在非音段的（超音段的，音韵的）特性，它们可能有或可能没有区别（音韵的）功能。与有特色的特征不同，它们不是固有的，而是累积的；因此，它们的存在只能通过与给定话语中的不存在形成的对比来检测。因此说，一个音节只有在与周围的非重音音节相比时才被视为是重音。换句话说，非音段的特征存在于句法的对比中，不是词形变化的对立。有几个这样的特点：（1）口音是通过响度或音高差异化赋予音节的突出性，通常两者兼而有之。（2）语调区别，例如，下雨了，不同于下雨了吗？（3）音量区别短音素和长音素（例如，在古典拉丁语中的元音和辅音。）（4）声调，通过在分段上相同音节不同的音高升降曲线，区别所谓的声调语言中的单词，诸如中文和许多非洲的方言。

Morphology. As the phoneme is a distinctive sound-class of allophones, so the morpheme is a form-class of allomorphs. But while the phoneme is without meaning, though t serves to build meaningful units, the morpheme is the smallest meaningful unit in a language. For example, the plural of the English noun is signaled by various allomorphs: /s/ in backs, /z/ in bags, /ez/ in boxes; less frequent is /en/ in oxen, children; in the last also the vowel changes, as it does, without additional suffix, in mouse, mice; sheep has a zero plural allomorph. There are also some borrowed plural allomorphs, as in data (which is in the process of becoming a singular) and phenomena and others (which are not), from Latin or Greek; cherubim from Hebrew; and alumnae and alumni from Latin (but formulae is often replaced by the Anglicized formulas). Most plural allomorphs must be attached, in English and in other Indo-European languages, to another morpheme that bears the lexical meaning. In some languages, however, this inflectional suffix appears as an orthographic vestige without phonetic realization, like the s in French tables. Morphemes that are lexical items—like bag—are termed “free”; those--like the plural -s—that are not lexical items are “bound’; but a single lexical item may consist entirely of bound morphemes, like re-sembled, in which ed signals the past, as in all weak verbs. Morphophonological variation occurs when different allomorphs are used in different phonological surroundings, such as English a before a consonant and an before a vowel. Sometimes it is useful to employ a morphophonological rather than a phonological transcription, together with rules that specify the proper phonological realization. A morphophonological transcription is indicated by curved brackets: French six, which may occur as /si/, /sis/, or /siz/, may be transcribed {sis}.

词法。由于音素是音位变体的独特声类，因此词素就是语素变体的形式类。但是，虽然音素没有意义，尽管t用于构建有意义的单位，但语素是语言中最小的有意义单位。例如，英语名词的复数形式是由不同的语素变体示意的：在backs中的/s/, bags的/z/, boxes 的/ez/ ；在oxen, children中不太常见的是/en/；最后，元音也变化了，就像在mouse, mice中没有附加的后缀一样；sheep有一个零复数的语素变体。也有一些借来的复数语素变体，来自拉丁语或希腊语的，在data（它在过程中成为了单数形式）和phenomena及其它中（它们不是）；来自希伯来语的cherubim；以及来自拉丁语的alumnae 和 alumni（但formulae常常被英语化的formulas所取代。在英语以及其它的印欧语言中，大多数的复数语素变体必须附加到另一个具有词汇意义的词素上。然而，在一些语言中，这种词尾变化的后缀表现为没有语音实现的正字法痕迹，像法语tables中的s一样。作为词汇项的词素—像bag—被称为“自由的”；那些不是词汇项的—像复数的-都是“附着的”；但单个的词汇项可能完全由附着语素组成，像re-sembled,其中的ed表示过去，就像所有的弱动词中的一样。当不同的语素变体用于不同的语音环境时，就会出现词素音位学的变异，例如英语位于辅音前的a和元音前的an. 有时，采用词素音位学而不是语音转录和指定合适的实现规则是有用的。词素音位学的转录由弯曲的圆括号来表示：法语six，可能以/si/, /sis/, or /siz/的形式出现，可能被抄录为{sis}。

Transformational grammar, omitting the phoneme stage of analysis and starting instead with a morphophonological statement, goes beyond mere allomorphs in that it reduces related forms to so-called underlying forms. These, it is claimed, the speaker has in his mental lexicon, either consciously or subconsciously, and from them he derives the surface lexical items (vertical lines are used to indicate these underlying forms for which no actual phonic realization is claimed). This |pugn| deliversimpugn and pugnacious; and |sign| delivers signal and sign; and in French|halt| delivers haut and altitude. But the surface forms may be so different phonological from one another that one wonders how a speaker could connect them at all, except because of their orthographical shape, or, in the French example, their etymology, from Latin altum. By the same token, one could establish an underlying form for both cow and bovine, namely, Proto-Indo-European *gwou- (an asterisk before a form indicates that it is reconstructed, not attested). From it both English words are ultimately derivable according to the sound laws of comparative linguistics. One must conclude that reality for the existence of underlying forms, though convenient for the linguist, cannot be claimed as a general condition typical of every user of language. But one does learn that reference to the mind, so vigorously repudiated by the anti-mentalist structuralists, can indeed, if employed without empirical controls, result in unscientific speculation. (The graphic similarity of impugn and pugnacious, however, has some merits not shared by a phonologically more realistic spelling, like impune. This fact ought to teach spelling reformers that orthographic complexities are possibly worth the semantic hints conveyed through morphophonological writing. But the chasm between cow and bovine remains unbridgeable.)

转换性语法，省略了分析的音素阶段，并以形态音位的陈述开始，超越了单纯的语素变体，因为它将相关形式归纳为所谓的基础形式。据称，这些说话者要么有意识，要么无意识地拥有自己的心理词典，而从其中，他导出表面词汇项（垂直线用来表示这些基础形式，因为这些形式并无声称的实际语音实现）。这个|pugn|传递了指责和好斗；而|sign|传递了信号和标志；而法语中的|halt|传递了高级和海拔。但表面形式可能在语音上相互如此不同，以至于人们对说话者如何将它们完全连接起来感到好奇，除了因为它们的正字法形状，或者在法语示例中，它们的词源来自拉丁语altum。出于相同的原因，人们可以为奶牛和牛建立一个基本形式，即原始印欧语系的*gwou-（一种形式前的星号表示它是重建的，未经证明的）。根据比较语言学的音韵规律，从中最终可引出两个英语单词。人们应该得出结论，现实存在着基础形式，虽然对语言学家来说很方便，但不能说成是每个典型的语言使用者的一般情况。但是，人们确实了解到，被反精神论结构语言学家如此极力否定的对心灵的引用，如果在没有经验控制的情况下使用，确实会导致不科学的推测。（然而，impugn 和 pugnacious的图形相似性有一些在语音上更现实的拼写所不具备的优点，像impune。这个事实应该使拼写改革者懂得，正字法的复杂性可能值得通过形态音位的写作来传达语义提示。但奶牛与牛之间的分歧依然无法弥合。

Syntax. The ordering of morphemes and words into larger units, into phrases and clauses and sentences, is the sphere of syntax (from Greek syntattein, “to put together”). Traditional Latin grammar has had, in terminology and in rules, a firmer hold on syntax than on the other levels of analysis. And the structuralist, overly concerned with the pieces that constitute a syntactic unit, neglected the rules of the larger syntactic arrangement of these units. Perhaps their most effective attempt was the Immediate Constituent (IC) analysis, practiced, since the 1940’s, by Rulon S. Wells and Eugene A. Nida and by Kenneth L. Pike, in his tagmemic theory. In a phrase like the queen of England’s husband, the hierarchical ordering of IC’s can be indicated by a system of parentheses: ([|the| |queen|] ; {|of| |England|}] [s] (husband). The remotely possible metaphorical meaning, “the queen of the husband of England” however, would look like this: (|the| |queen|) ([of] [{|England| |s|} {husband}]).

句法。将词素和单词排序成更大的单位，排序成短语、从句和句子，是句法的范围（来自希腊语syntattein，“放在一起”）。在术语和规则方面，传统的拉丁语语法对句法具有一种比其它分析层面更牢固的把控。而结构语言学家，过度关注构成句法单位的片段，却忽视了这些单位更大句法安排的规则。也许他们最有效的尝试是直接成分（IC）分析，自20世纪40年代以来，由鲁伦·S.威尔斯和尤金·A. 奈达，以及由肯尼恩·L. 派克在其标记语言学理论中的实践。在短语中，例如英格兰女王的丈夫，直接成分的分等级排序是由括号系统表示的：([|the| |queen|] ; {|of| |England|}] [s] (husband)。然而，略有可能的隐喻含义“英格兰丈夫的女王”看起来像这样：(|the| |queen|) ([of] [{|England| |s|} {husband}])。

In the mid-1950’s, transformational linguists began to put syntactic constructs not at the end but at the beginning of linguistic analysis. In a way, IC’s continued to be involved (though now their hierarchy was often visualized by a tree design) because a syntactic rule must state what constituents of a sentence belong together to form syntactic subunits, in what sequence they do this, and in what hierarchical order. While structuralism starts with the smallest analytical units (allophones and phonemes) and builds up a sentence, transformational grammar begins with the sentence, indeed with an underlying deep structure that, as it were, states the theme and general meaning of the message, upon which is mapped, through sets of rules, the phonetic realization, or surface structure. These divergent attitudes are based on and, in turn, lead to divergent philosophic attitudes toward language and linguistics: structuralism is inclined toward empiricism and stresses the importance of observable facts; transformational grammar, toward rationalism and intuition. This dichotomy is as old as philosophy itself: in antiquity it is paralleled by Aristotelianism and Platonism, respectively.

在20世纪50年代中叶，转换性语言学家开始将句法结构放在语言学分析的开始，而不是放在末尾。在某种程度上，直接成分继续参与其中（虽然现在它们的层次结构通常被一个树状设计可视化），因为句法规则必须说明，句子的什么成分，以什么样的顺序，以及按什么层级秩序一起构成了句法子单元。而结构主义从最小的分析单元开始（语位变体和音素）并建立起句子，转换性语法从句子开始，事实上具有潜在的深层结构，可以说是陈述了信息的主题和一般含义，通过一套规则，语音的现实，或表面结构来映射信息。基于这些不同的态度，反过来导致了对语言和语言学的不同哲学态度：结构主义倾向于经验主义，并强调看得见事实的重要性；转换性语法趋向于理性主义和直觉。这种二分法与哲学本身一样古老：在古代，它分别与亚里士多德学说和柏拉图学说并行。

Lexicon. The words of a language constitute its lexicon. A “word” is notoriously difficult to define linguistically and as one language easily appropriates a word from another (loanword), linguists use it as a pragmatic rather than theoretical unit. Lexicographers record, sometimes translate, words in descriptive or historical (etymological) dictionaries.

词典。一种语言的单词构成了其词典。在语言学方面，一个“单词”难以定义是人所共知的，而且作为一种语言会轻松地从另一种语言中挪用一个词（外来语），语言学家将它用作一种实用的，而非理论的单元。在描述性或历史词典（词源的）中的单词时，词典编纂者记录，有时进行翻译。

Diachronic Linguistics. Much that pertains to this subject has already been said in this article, because modern linguistics began as a historical science and remained so exclusively for several decades; and the terminology given for descriptive linguistics is valid also for historical linguistics. In the main, current diachronic linguistics still honors the aims and methods of the neo-grammarians. The structuralist reform, however, caused linguistic changes to be no longer viewed atomically—for example, by considering the history of a single sound from Proto-Indo-European to modern Spanish—but as pertaining, at a given historical moment, to the prevailing linguistic structure as a whole, and to be stated and explained as an integral part of the economy of the system. Diachrony concerns itself chiefly with the past. But it is not impossible to discern apparently sporadic present drifts that, in the light of the linguist’s experience, bid fair to become standardized in the future—for example, the loss of the THIRD person singular ending -s in the English verb. But no linguist will venture sure predictions of whether and when this will actually happen. DRIFT

历时语言学。本文中已说过许多与该主题有关的内容，因为现代语言学开始作为一门历史学科几十年来一直如此；而且描述性语言学给定的术语也适用于历史语言学。一般来说，当前的历时语言学依然尊重新语法学家的目标和方法。然而，结构主义的改革引发了语言学的变化，不再以原子方式来看待—例如，通过考虑从原始印欧语系到现代西班牙语单一声音的历史，但在给定的历史时刻，它与作为一个整体的流行语言的结构有关，并作为系统经济不可或缺的一部分有待被陈述和解释。历时分析主要关注的是它自身的过去。但根据语言学家的经验，辨别出明显零星的当前漂移并非不可能，有可能成为未来的标准—例如，在英语动词中第三人称单数结尾-s的丢失。但没有语言学家会冒险准确地预测这种情况是否以及何时会真的发生。

The efforts of transformational grammar in the diachronic domain differ somewhat in technique, but they cannot deliver greatly divergent results. However, since the transformationist’s historical underlying forms may be, just like his synchronic underlying forms, pure constructs and logical abstractions without an intrinsic requirement for historic reality or near-reality, they are under no constraint to correspond to, or even resemble, anything that ever existed. These historical forms are therefore of questionable historical validity.

在历时领域中转换性语法的努力在技巧上有所不同，但它们无法提供截然不同的结果。然而，由于转化论者历史的基础形式可能是，就像他的共时性基础形式，纯粹的结构和逻辑抽象一样，没有对历史现实或近乎现实的内在要求，它们不受任何约束，甚至无法与过去存在的任何相似的东西相对应。因此，这些历史形式的历史有效性是值得怀疑的。

Dialectology. In linguistics, the term “dialect” does not necessarily refer to a substandard or socially inferior type of speech. Indeed to the linguist the distinction between language and dialect is superfluous. However, in accordance with common usage, he may use the term “language” to designate a standard or literary form of speech used in an area, often a nation, where various local and social dialects exist, and where speakers are depending on their education, mono-dialectal in either the local or the standard dialect, or bidialectal. But there is still no reason for not using the term “standard dialect” rather than “standard language.” Social prestige is often accorded the standard dialect because it is the vehicle of most literature and writing in general, hence associated with learning and education. Nonstandard speeches, on the other hand, bear a stigma because they usually betray a lack of schooling in their users. From a purely linguistic, nonsocial point of view, no dialect is superior to another.

方言学。在语言学中，“方言”一词不一定是指不合规格或社会上低人一等的言语类型。实际上，对于语言学家来说，语言和方言之间的区别是多余的。然而，按照通常的用法，他可以使用“语言”一词来指定一个地区所使用言语的标准或文学形式，常常是一个存在各种地方和社会方言的国家，那里的说话者依赖于他们的教育，要么是当地方言，要么是标准方言的单一方言，或者精通两种方言。但是，依然没有理由不使用“标准方言”，而不是“标准语言”一词。社会声望通常被赋予标准方言，因为，一般来说它是大多数文学和写作的载体，因此，与学习和教育有关。另一方面，非标准的言语带有污名，因为它们通常泄露了它们的使用者缺乏学校教育。从纯粹的语言学，非社会的视角来看，一种方言与另一种方言并无优劣之分。

It is difficult to define the term “dialect,” and to determine where a dialect boundary should be drawn on a map, because a dialect is not a thing that occurs in nature. It is, rather, a class of such idiolects (an idiolect is the manner of speaking of an individual) as are, for good reasons, appropriately gathered under one heading. But even if one ignores the fact that “idiolect” is already to a degree an abstract term and that individual speakers do not always use the same idiolect, one still must state the criteria by which idiolects are assigned to one dialect. The linguist has no quantitative criteria at his disposal, no scales of measurement that furnish an answer. He cannot affirm categorically what kind and what number of traits—phonological, morphological, syntactical, lexical—two or more idiolects must share to be classified in the same dialect, and to what degree they must differ to be assigned to different dialects. Mutual intelligibility is certainly a requirement. But intelligibility itself is a matter of degree: does it prevail when the general meaning is understood, or something in between? Indeed one may for sound linguistic reasons choose to put mutually “intelligible” idiolects into completely different dialects.

想定义“方言”一词，并且应该将一门方言的边界画在地图的哪里很难，因为方言并不是自然界发生的事情。相反，它是这样一类个人习语（个人习语是个体的说话方式），出于充分的理由它们被适当地聚集在一个标题下。但即使人们忽略了这样一个事实，即“个人习语”在某种程度上已经是一个抽象术语，而个别的说话者并不总是使用相同的个人习语，人们依然必须说明将个人习语归属于一种方言的标准。语言学家并没有可使用的量化标准，没有提供任何答案的测量尺度。他无法明确地肯定什么种类和数量的特征—音韵学、形态学、句法、词法—两个或以上的个人习语必须共享才能归为同一方言，以及它们必须达到什么程度的不同才能被归属于不同的方言。相互的可理解性无疑是一种必要条件。但可理解性本身是一个程度问题：当理解了一般含义，或介于两者间时，它会具有优势吗？实际上，出于合理的语言原因，人们可能选择将相互“可理解的”个人习语归入完全不同的方言中。

Of course, obvious cases are dealt with easily, sometimes on geographic or national grounds. A speaker of Parisian French and one of Roman Italian speak palpably different dialects, and so do a speaker of Roman and one of Milanese Italian, even though all these dialects are historically and typologically related. They are all modern forms (descendants”) of Latin, and they are all subsumable in the larger class of the Romance dialects, or the Romance diasystem. But whether or not an inhabitant of village X and one of village Y only a few miles distant speak the “same” dialect may pose a difficult problem. In fact, there exists, at the present stage of linguistics, no substitute for the good sense and the experience, in short, for the art, of the linguist. Thus disputes on dialect boundaries and the number of dialects to be counted in a given area are likely to be, not over facts, but over opinions concerning how to categorize the facts.

当然，显而易见的情况很容易处理，有时是基于地理或国家的理由。一个说巴黎法语的人和一位说罗马意大利语的人说着明显不同的方言，而一个说罗马语的人与一位说米兰意大利语的人也是如此，尽管所有这些方言从历史的观点和从类型学角度都是相关的。它们都是拉丁语的现代形式（衍生物），而且它们都是可归入罗曼语的方言，或更大的方言系统类别。但是，无论X村的居民与仅隔几英里的Y村居民是否说“相同”的方言都可能引发难题。实际上，在语言学的现阶段，并没有出现语言学家良好感觉和经验的替代品，简言之，是为艺术。因此，有关方言边界和在给定区域计算的方言数量的争论很可能不是关于事实，而是关于如何对事实进行分类的意见。

Besides local, or horizontal, dialects one must also recognize social, or vertical, ones. In the latter, several dialect layers occupy the same geographic location. Social dialects have especially attracted the attention of linguists in the second half of the 20^th century.

除了地方的，或同一层次的（横向的）方言，人们必须识别社会方言或不同层次的（纵向的）方言。在后者中，数个方言层占据了相同的地理位置。在20世纪下半叶，社会方言尤其吸引了语言学家的注意。

The gathering of data on local dialects began in Europe, coinciding with the growing interest in synchronic linguistics. In 1881 the German linguist Georg Wenker published six maps that started an atlas of German dialects. Since then, dialect atlases for various countries, such as France, Italy, and the United States, and for smaller regions have been designed, and work has continued in numerous places. Generally these atlases consist of a number of outline maps of the area under investigation. On them are marked—by a number rather than by name—all the localities whose speech was examined by specially trained, skilled linguistic fieldworkers, who do their work either by recording and transcribing conversations or with the aid of a questionnaire that solicits native responses, or by both methods combined. Each map in the atlas normally deals with one word or one short phrase, whose shape is entered in phonetic transcription at each of the localities shown on the map. Thus the reader can see how a given word is pronounced in hundreds of places (nearly 1,000 of them, for example, in the dialect atlas of Italy published in the 1930’s) or by what different names a give object is called, or how phrases of the same meaning are worded. Obviously an atlas of this kind provides only the phonetic and lexical raw material for the linguist, who then starts to analyze and classify the data. One of his tasks is to determine where and how one may reasonably draw dialect boundaries on any one or a whole set of such maps. Usually these boundaries coincide with bundles of isoglosses. An isogloss is a line that separates two adjoining but different linguistic usages. An example from American phonology is the pronunciation of greasy with /s/ or /z/, and one from the lexicon is the use of sack versus bag. In addition to providing linguistic information, such a dialect map lends itself to historical, socioeconomic, and ethnic interpretations, since linguistic data may betray nonlinguistic conditions of both the past and the present.

对当地方言的数据收集始于欧洲，与人们对共时语言学日益增长的兴趣不谋而合。1881年，德国语言学家格奥尔格·温克出版了六幅地图，开启了德国方言的地图册。从那以后，方言地图册在各国出现，比如法国、意大利和美国，而对于较小的地区也进行了设计，而且许多地方的工作仍在继续。通常，这些地图册由一些被调查区域的轮廓图组成。在地图上--用编号，而不是用名称--标明那些言语的所有聚居地由受过专门训练，技能娴熟的语言学实地工作者来调查，他们的工作要么通过记录和转录对话，或借助于征求母语回应的调查问卷，或通过两者结合的方法来进行他们的工作。地图册中的每张图通常涉及一个单词或一个简短的短语，在地图上显示的每个地点中用语音学音标输入其形状。因此，读者可以看到一个给定的单词在数百个地方（例如，在20世纪30年代出版的意大利方言地图册中，有近1000个）是如何发音的，或者用什么不同的名称称呼给定的对象，或者相同含义的短语是如何措词的。很明显，这种地图册为语言学家只提供了语音和词汇的原材料，然后语言学家开始对数据进行分析和分类。他的任务之一是确定在何处以及如何合理地在任何一张或整套此类地图上画出方言的边界。通常这些边界与同言线束相一致。同言线是分离两种邻接，但不同语言用法的线。美国音韵学的一个例子是greasy的发音为/s/ 或/z/，词典的一个例子是使用sack与bag相比。除了提供语言学的信息外，这种方言地图适合于历史的、社会经济学的，以及种族的解释，因为语言学的数据可能会违背过去和现在的非语言条件。