The Syntax-Semantics-Prosody Interface in Legislative Language

Though Chinese and English legislative language share similarities in certain aspects, they contrast most sharply in their syntactic structures and representations, owing to a number of social, cultural and typological factors. Traditional works have directed most of their attention to the lexicon, i.e. the collection of technical terms which are indispensable to the working of law, and the analysis of idiosyncrasies of those terms. Not much effort, if any, has been at present devoted to the study of the syntactic properties of legislative language and the way syntactic computation interfaces with semantic layer, which, the author believes, play a more determinate role in the constructing and interpreting of legislative clauses. This paper first attempts to examine the syntax-semantics interface in Chinese and English legal clauses from the levels of recursion, locality of reference and the general principle of economy respectively, and then tries to study how the position and stress of a word affect the syntax-prosody interface and the information flow. In the end, it concludes that syntax is responsible for base meaning and while prosody highlights a part or parts of that meaning.


Introduction
Linguistic analysis of the language of the law has been gaining more momentum in recent years both at home and abroad.The 20 th century is deemed as the age of "linguistic imperialism" by the western academia (Song, 2009: p. 152).Language is no longer treated as a vehicle or means for the expression of ideas and concepts.Instead, it is discovered that language itself contributes much to what that concept or idea is.The same perception has been accepted in the domain of jurisprudence as well, i.e. legal clauses or statutes are regarded not as the mere bearers of the law but they are deemed as the very makers of the law."The last two decades have witnessed what has come to be known as the linguistic turn in a number of disciplines" and the study of law is one of them (Salmi-Tolonen, 2004: p. 1167).She echoed John Searle by agreeing that "social reality, including law, is rooted in language".Among the early endeavors taken in this field is the classic work of D. Mellinkoff, initially published in 1963, it has provided the most comprehensive and detailed description of legal language by far.Not unexpectedly, this book attaches a high priority to language rather than to law.And with the continuation of the tradition of digging into the etymology of legal lexicon to explain word features such as archaism, formalism and ambiguity etc. which, however, will not be a focus in the present paper, the language of the law therein is commented upon as wordy, unclear, pompous and dull, which might be viewed as general statements about both of its syntactic and semantic features but it goes not much further.
Other modern linguistic efforts made in the analysis of legislative language can be summarized as mostly semantic or pragmatic accounts exploring its interpretive possibility or ambiguity at various levels.With regard to the comprehension of a statute, different scholars usually take different approaches.In general, there are two interpretive camps: textualists, who confine themselves to the legal text itself for an answer, and contexualists, who search beyond the text or "the four corners of law", and to look at a law's history for an accurate and proper application of the law (Brunson, 2012: p. 1).Though he admits that textualists often have to surrender to their opponents and take up a contextual position in certain cases.In fact, quite many forensic linguists prefer a more dynamic and contextual approach, for example, H. L. A. Hart (1961) resorts to J. L. Austin's speech act theory for a definition of legal language, asserting that "the primary function of (legal) words is not to stand for or describe anything but a distinct function" (Goodrich, 1984: p. 523).And "Legal rules concerning interpretation are an assorted mixture of conventions, maxims, principles and authorities" (Hutton, 2009: p. 64).With reliance on M. A. K. Halliday's functional theory of language, A. Johnson and M. Coulthard (2010) propose that meaningmaking is embedded in social practice and all sorts of complicated contexts, and state that.
"Context is dynamic and socially constructed through and by discourse-both in its linguistic and non-linguistic semiotic modes-and we know that the legal world is context-rich.It is peopled by a hierarchical mini-nation of judges, lawyers, police and law-enforcement officers and then the common man and woman… Its texts are also richly layered with meaning; its language has evolved over many centuries and its peculiar form is a result of this history and specialized use." (Johnson & Coulthard, 2010: p. 1) Although these studies have primarily inquired into either textual or contextual way to interpret a statute, neither approach has paid sufficient attention to the syntactic construction of that statute.From the perspective of formal linguistics, the structure, or more specifically, the position of a word and the way it interacts with its surrounding words contribute more to the determination of its meaning.The present paper aims at unwrapping some of the intertwinement incurred when syntactic projections map on to LF (semantic level) or PF (phonological level) respectively and proposes that the structural peculiarities of legal clauses frame and determine in most parts the outcome of their interpretation.

The Syntactic Features of English Statutes
Legal English occupies a unique position in the history of English writing.Scholars have offered a wide array of choices for its description and explanation.Apart from its distinct choice over technical terminology, one other most prominent feature it owns is its syntactic makeup which can be described as long, complex sentences with "intricate patterns of coordination and subordination, and is marked by repetition, a complete absence of color and humanity, and has a preference for passive voice" (Maley, 1987: p. 25).G. Venturi, who works by collecting and comparing chunking corpora of Italian and English legal texts with corpora of ordinary language, has cited a similar linguistic study conducted by Garavelli, who has found that legal language is "different but not drastically independent from ordinary speech" and rather it is "a special use of lexical and syntactic particularities characteristic of common language and it can be seen both as an extension and as a reduction of the possibility offered by ordinary language" (Venturi, 2010: p. 4).There are those who trace back to the origin of English law for the reason why law language is what it is.The core of English legal language is believed to be "medieval French of a particular type… the technical language of the French speaking lawyers in the England of Edward I and later" and evidence also shows that the law of that era was in general written in French (Woodbine, 1943: p. 395).Therefore, a study of the Medieval French syntax and legal lexicon may help to shed some light upon English legal sentences.
In the following sections, three aspects of the syntactic features of a statute will be examined respectively.They are what is the cause and formation of its unusual lengthy structure, and how one move from a word within the containment of that statute will bring changes to its structure and meaning alike, and how a particular syntactic structure interfaces with both semantics and phonetics.Finally, the development of English legal structure will be tackled with in some details.

Recursion as the Means to Generate Complex Sentences
Recursion is seen as one of the essential nature of human language, i.e. a limited number of words or phrases can be reconstructed to form hierarchical structures by ways of self-embedding, which can produce a potentially unlimited number of different sentences, and sentences of unlimited length (Dikken & Lahne, 2013: p. 655).According to MP (Minimalist Program), the derivation of a sentence has been reduced to two steps: merge and move.Merge is an external operation incorporating new elements such as adjuncts, coordinate or subordinate constructs into sentences, while move is an internal operation which affects movement of elements such as whwords within the sentence showing a syntactic dependencies abided by locality restrictions (Chomsky, 1995).
In legal documents, it seems that merge is the much more prevalent operation than move is because the syntax of legislative writing is invariably marked more by the addition of coordination and subordination than by the movement of wh-words which is generally responsible for the formation of interrogative sentences.Among the four types of English sentences: declarative, imperative, exclamative and interrogative, declarative is generally preferred to the exclusion of the other three types in legislative writings (Xiao, 2007: p. 179).For example: "1) When the justices of the peace of any 'county assembled at quarter session' have agreed that the ordinary officers appointed for the preservation of the peace are not sufficient for the preservation of the peace, the protection of the inhabitants, and the security of property within the county, and when they have set forth the same, and have declared how many constables are in their opinions needed for the aforesaid purpose, and what rates of payment it would be expedient to pay to the chief and other constables, and when such report has been sent to one of Her Majesty's Principal Secretaries of State.and when one of Her Majesty's Principal Secretaries of State has made and finally settled any rules for the government, pay, clothing, accountrements, and necessaries of such constables, and when such rules as finally settled (?) have been received from the Secretary of State by the clerk of the peace of such county; The SAID JUSTICES may appoint a chief constable, & c." (Coode, 1845: pp. 64-65) Here the underlined "when" works as the conditional "if" and it is not a true wh-word with the purpose of time-asking.Hence no wh-movement is involved in any of the "when" clauses.In the above 166-word article, there are five "when" conditional clauses with one main clause.All these clauses are subordinate to the main clause and are connected to the latter by merging as its adjuncts.While the five "when" clauses form a parallel relation with each other by coordinating through an "and".As asserted by generative grammar, adjuncts could be merged more than once to the same VP.This structural information provides the typical formula for a statute: "If P1 + P2, then Q" (Zhang, 2005: p. 198).
With respect to Chinese written legislative language, the Chinese statutes are characterized by the prevalent use of imperative and declarative sentences (Zhang, 2005: p. 198).The following are examples using imperative sentences.
The existence of a large amount of imperative sentences in Chinese statutes reveals another big difference between Chinese syntax and English syntax.It is explained that English belongs together with French and German to the non-pro-drop languages, while Chinese, Japanese and Slavic languages are classified as pro-drop languages which usually refer to a language in which certain classes of pronouns may be omitted when they are in some sense pragmatically inferable (Wikipedia).In more familiar words, English is often labeled as a subject prominent language while Chinese as a topic prominent language.As for the Chinese, the omission of subject is a common practice while it is an impossibility for the English.Please compare the English version of example ( 2) and (3), with (2) corresponding to (4) and (3) to (5) respectively.4) Article 2. A marriage system based on freedom, monogamy and equality between man and woman shall be implemented.(Marriage Law, Article 2) 5) Article 21, para.4. It shall be forbidden to drown or desert infants or commit any other act of killing infants.(Marriage Law, Article 21, para.4) From Articles (4) & ( 5), it can be observed that the need for a subject is a syntactic requirement which must be satisfied if the English sentence is to be judged grammatical.Moreover, there is an apparent difference between the subject of (4) and that of (5).In (4) "a marriage system" is a concrete nominal phrase or DP while in (5) "it" is an expletive with no semantic but only a functional role to play.The rule here is the application of EPP which hypothesizes that subjects are obligatory (Chomsky, 1981).Hence, in the absence of a real subject, an expletive must be inserted.As for the reason why null subject is allowed in the Chinese Articles ( 2) and (3), there are two accounts.One is within the pro-drop language type there exists a subtype tagged null subject languages whose grammar permits an independent clause to lack an explicit subject, such a clause is then said to have a null subject (Wikipedia).Chinese language is a member of such null subject types, and therefore the disappearance of subjects from the Articles ( 2) & ( 3) can be elucidated in some way.The second account is that imperative sentences typically work without subject.For English imperative sentences, the omitted subject is invariably the implicit second person "you" (Quirk et al., 1985: p. 828), thus, no semantic ambiguity will arise from such structures.However, Chinese imperative sentences are more complicated.The omitted subject is not always "you".In ( 2) and (3) the implicit subject is definitely not "you".And a possible subject for (2) would be "this country" or "this government" and "the law" for (3).

Syntactic Ambiguity and Locality of Reference
What is most contentious in the field of jurisprudence concerns the way the law is interpreted.In the process of which many factors such as morality, social justice, political needs other than the language of the law itself come into play.O. W. Holmes, an early exponent of "legal realism", who is quoted as saying that "The life of the law has not been logic: it has been experience" (Hutton, 2009: p. 11).The opposite view, called "legal positivism", believes law is autonomous and "It is necessary to distinguish between law as it is and law as it should be" (Hart, 1983).
Putting aside non-linguistic considerations, the present paper proposes that even the language used in the writing of a statute can be enough source of misinterpretation.There exist eight types of syntactic structures that either bring about ambiguity or lead to total failure in understanding.They are 1, the use of excessive complex sentence (multiple embedded sentences); 2, the overuse of passives; 3 & 4, whiz-deletion and unclear pronoun reference; 5, nominalization; 6, multiple negation; 7, archaic and misplaced prepositional phrases; 8, legal language's own set of articles and demonstrative pronouns (Crandall & Charrow, 1990: pp. 6-10).The following example is to illustrate 1 and 3 & 4 respectively.6) "You must never speculate to be true any insinuation suggested by a question asked a witness."(Crandall & Charrow, 1990: p. 6) 7) "In determining the weight to be given such opinion, you should consider the qualifications and credibility of the expert and the reasons given for his opinion.(Crandall & Charrow, 1990: p. 7) The underlines are added by the author.The example (6) proves that the more embedding a sentence has, the harder it is for it to be understood.The example (7) suggests that if "given" is replaced by "he gives", the sentence will be much clearer.The underlined "his" will have an antecedent "he" within its local domain.And according to Binding Principle B, i.e. a pronoun must be free in its local domain, it then can be deduced that "he" can refer to "the expert", which is not in its local domain.
Crandall & Charrow are not the only ones who have noticed the troubles caused by syntactic structures favored in most of the legal writings.Other types of syntactic or grammatical ambiguity have been found.They are the following: 1, the alternation between "and" and "or", 2, the ambiguity of plurals, 3, scope ambiguity which often depends on the so-called "the last antecedent rule" for a solution, 4, the ambiguity caused by non-restrictive modification construction, 5, underspecification and aspectual ambiguity (Vogel, 2009: pp. 27-42).Take the following example for an illustration of type 3.
8) The possibility of admission is available to girls and boys under six years of age.(Vogel, 2009: p. 29) The underline is added by the author.To interpret the underlined ambiguous parts, the following rules are used.(Solan, 1993: p. 29) Then the following interpretation will be obtained.Generally speaking, ( 9) is the unmarked reading.9) ((girls )and (boys) under six years of age.) 10) (girls and ((boys) under six years of age.))

The Syntax-Semantics Interface in Complex Legal Sentences
As stated in the preceding parts, legislative language is marked by syntactically involved sentences.Using a parser based on a context free grammar, the following model of legal sentences has been developed in order to illustrate the complexity of legal sentences.
Figure 1 has been constructed with data from Dutch statutes.When compared with English legal language, it is found that they share many similar syntactic features.The English legal syntax is characterized by the use of declarative, conditional clauses, parallel and especially prolonged sentences (Ji, 1998: p. 32).For example: 11) With respect to any existing internal tax which 1 is inconsistent with the provisions of paragraph 2, but which 2 is specifically authorized under a trade agreement, in force on April 10, 1947, in which 3 the import duty on the taxed product is bound against increase, the contracting party imposing the tax shall be free to postpone the application of the provisions of paragraph 2 to such tax until such time as it can obtain release from the obligations of such trade agreement in order to permit the increase of such duty to the extent necessary to compensate for the elimination of the protective element of the tax (GATT, 1974).
The above Article consists of one long sentence with 100 words.In accordance with Figure 1, this sentence contains a conditional clause which begins with "with respect to" and the main clause which begins with "the contracting party".If analyzed in terms of syntax, it contains one fronted adverbial clause attached to the main clause.This one adverbial clause itself is a complex one with three embedded attributive clauses in it, while the main clause is also complicated since it has adverbials phrases like "such as" and "in order to" to prolong the structure further.
Based on a RRG (Role and Reference Grammar) perspective, it is proposed that in a complex construction there are nexus relations, or the syntactic relations between the units (Van Valin, 2005: p. 183).He divides complex sentences into three main types: type 1, coordination = unit 1 + unit 2; type 2, subordination = unit 1 embedded in unit 2; type 3, cosurbordination = (unit 1 + unit 2) these two units together embedded in one sen- tence.Based on the syntax-semantics interface model provided by R. D. Van Valin, a simplified analysis of example (11) has been drawn below.
Figure 2 illustrates that syntax and semantics interface with each other bilaterally.It has been pointed out that the relation between semantics and syntactic structure is not determined unilaterally in either direction, rather it is the way semantics and syntax interface mutually that determines how one maps on to the other (Mei, 2008: p. 409).For legislative language, its syntactic structure also works actively with semantics towards an integrated meaning.Legal syntactic structure may in great many parts how a legislative sentence is interpreted.

Legal Syntax and the Principle of Economy
Legal documents are best known for its wordy, arcane and uncommunicative style.D. Mellinkoff has long urged "the plain English movement" to promote the simplification of English legal language.Then the question is whether it is really possible or necessary to replace the traditional lengthy style with shorter syntax and more ordinary words?The generative linguistics has long observed the principle of economy, also known as Occam's razor, which requires that among competing hypotheses, the one with the fewest assumptions should be selected (Wikipedia).The principle of economy is manifested in the economy of derivation and representation.It is stated that the levels of representation should be reduced to as few levels as possible, which requires the elimination of superfluous elements, while the economy of derivation follows a least effort condition, i.e. "when there are multiple possible derivations all deriving the same representational output, economy conditions compel the grammar to perform the most economical one" (Epstein et al., 2013: p. 493).
Although the principle of economy is more of a technical guideline to the computational process of a sentence, it offers helpful hints to the question posed above, i.e. should the formation of legislative clauses obey this general principle as well?In such a case, the peculiarities of legislative language have to be considered in order to make a choice between a simple style and a complex one.The essential features of legislative rules are described as clarity, precision, unambiguity and inclusiveness (Bhatia, 2010: p. 38).The first three features are somewhat in conflict with the last one because precision is not compatible with all-inclusiveness.If precision is aimed at, it is only very natural that longer sentences are preferred over shorter one for longer sentences would include more details and information.And this might partially account for the prevalence of lengthy sentences in legal documents.Suppose now that long sentences are to be replaced with short ones and at the same time diffi- cult words are to be substituted by common words, the resulting effect might be a double-edged sword.The apparent advantage is that it might ease understanding and facilitate communication.The disadvantage, however, is that the language used appears to be informal and consequently lose the solemnity of law.After all, law serves as a deterrent to potential wrongdoers in a society and hence the more formal the language of the law is, the more power and control it possesses.Secondly, when judges deliver sentences, the decree versed in formal terms seems to be more forceful and authoritative than those in plain words.And this latter consideration will bring another element, i.e. the phonological aspect of law, under scrutiny in the next section.

The Syntax-Prosody Interface in Legislative Language
In recent years, quite a few researchers direct their attention to how the way people speak might influence the working of syntax.Of course, there are opponents who believe in the autonomy of syntactic computation.But more researchers begin to realize syntax works with prosody to some extent.There is a doctrine which labels syntax and prosody as two separate generative systems: one is syntactic phrase marker and the other is prosodic structure, which, under mapping constraints, "forms the set of possible pairs <s, p> corresponding to the wellformed sentences of the language" (Buring, 2013: p. 860).And he further suggests that prosody does influence syntax since there are ample cases where prosody affects word and constituent order such as heavy NP shift and prosodic extraposition.Moreover, prosody also interacts with information structure to provide the contrast between giveness and focus.The notion "prosodic connotation", which defines prosody itself as a phonological means to express the speaker's mood and therefore carries with it symbolic and stylistic meaning, is also emphasized (Feng, 2007: p. 56).
In legislative language, prosody plays a part too.As stated above, the use of formal style may pass along messages of power and solemnity, which may strengthen the law's regulating and judicial role in a society.Besides, the court decree, when delivered in formal terms, sounds much more forceful than it is said in informal terms.Hence, formal style should not be sacrificed too quickly for the sake of easier communication.As for how prosody may influence syntax and subsequently semantics, take the passive construction for example.
12) The aforesaid representations were false and were then and there known by defendant to be false.(Crandall & Charrow, 1990: p. 6) The underline is added by the author.Because the use of passive construction, the focus, which should be on "defendant" has been falsely shifted to "representation".Here prosody is meant to bring forth the information structure.The syntax-prosody mapping includes two types of constraints: narrow syntactic mapping and extraneous feature mapping which will not be elaborated here.

Conclusion
This paper has explored the syntax-semantics-prosody interface in legislative language.The syntactic features of English legal clauses are marked by their unusual length, complexity and the declarative mood they customarily resort to, while Chinese legal clauses are characterized by their declarative and imperative mood they adopt.To explain the reasons why such syntactic structures are favored and also how those particular structures work with semantic layer, recursion, locality of reference and the principle of economy have been examined one by one.Recursion is basic means to the formation of complex structures like embedding and coordinating; locality of reference helps anaphora and pronouns to find their antecedents; and the principle of economy explains that though it seems on first sight that lengthy legal sentences disobey that universal principle by using structures more than what is required by communicative needs, they are in fact the compromise of many intertwining factors like social, judicial and professional considerations.And the fact is they do obey economic principle in a complicated way.Finally, how prosody influences syntax is examined and prosody can also have an effect on the information structure in order to give different focuses, though much work is still needed in this field.