-B-

Bach-Peters paradox
SYNTAX: paradox in the description of sentences such as (i), first noted by Emmon Bach and Stanley Peters.
(i) [the student who deserves iti]j will get [the reward hej works for ]i
If iti is intended to be co-referential with the reward hej works for, and hej is intended to be co-referential with the student who deserves iti, and if the corefering terms are equated in the description, we have the paradox that a term a which properly contains a term b, is equal to a term b which is properly contained in a (the paradox being that a term must be both equal and unequal to another term). In the case of (i) the paradox is avoided if the description is something like (ii).
(ii) for all x, x:a student & for all y, y:a reward (if x works for y & x deserves y, 
     then x will get y)
LIT. Bach (1970, May (1985).

Back
PHONOLOGY: a °feature which characterizes sounds that are produced by retracting the body of the tongue from its neutral position. EXAMPLE: the vowel [u] differs from [i] in that [u] is characterized by [+back] and [i] by [-back].
LIT. Chomsky & Halle (1968), Halle & Clements (1983).

Back-formation
MORPHOLOGY: a type of word formation by °analogy. Back-formation occurs when speakers of a language assign a regular derivational structure to a word, although a part of this structure, namely the °base, did not previously exist. If this new base becomes a word of the language, it is called a back-formation. EXAMPLE: speakers of English have reinterpreted the °primary compound baby-sitter as being a °synthetic compound, i.e. consisting of the base baby-sit and the suffix -er, and on the basis of this noun they have coined the verb to baby-sit. Another example is self-destruct from self-destruction. In the literature, the existence of back-formation is taken as evidence for a °word-based morphology.
LIT. Marchand (1969), Aronoff (1978), Scalise (1984), Booij (1989), Spencer (1991).

Backward anaphora
°Cataphoric relation.

Bahuvrihi compound
MORPHOLOGY: a term first used by the Sanskrit grammarians to refer to a particular type of compound, viz. compounds that lack a phonologically visible head, and which predominantly refer to pejorative properties of human beings. EXAMPLE: in regular compounds such as bloedneus 'bloody nose' the word neus functions as the head, and bloed as the modifier which attributes a property to the head neus, viz. that it is bleeding. Bahuvrihi compounds like wijsneus 'wise guy' (lit. 'wise nose') or roodhuid 'redskin' (='Native American') are crucially different in this respect. A wijsneus is not a nose that is wise. To distinguish between both types of compounds, the term °endocentric compound is used for compounds such as bloedneus, while the term bahuvrihi compound or exocentric compound is used for compounds such as wijsneus.
LIT. Kiparsky (1982), Spencer (1991).

Bare plural
SEMANTICS: a plural noun phrase without an overt determiner. EXAMPLE: cats, or pretty girls with blond hair are bare plurals. Bare plurals can have a °generic interpretation (in (i)) or an °existential interpretation (in (ii)):

(i)  Cats are intelligent
(ii) Cats ruined my garden
Carlson (1977) analyzes bare plurals as basically names for kinds.
LIT. Carlson (1977).

Barrier
SYNTAX: a category which serves as a blockade for °government and/or °movement. The original definition (from Chomsky 1986b) is:

(i) g is a barrier for b iff (a) or (b):

    a. g immediately dominates d, d a °Blocking Category for b
    b. g is a Blocking Category for b, g =/= IP
LIT. Chomsky (1986b), Rizzi (1990), Lasnik & Saito (1991).

Base
MORPHOLOGY: a term used for that part of a morphologically complex word to which derivational or inflectional morphemes are added. EXAMPLE: in the English words fatherhood and fathers, the noun father functions as the base to which the derivational suffix -hood, and the inflectional suffix -s, respectively, are added.

Base component
SYNTAX: Module of the grammar in which °D-structures are generated by means of °phrase structure rules and the °Projection Principle, on the basis of information from the °lexicon.

Base-generated
SYNTAX: Whatever is an element of a °D-structure, hence is generated by the °base component.

Belief sentence
°Propositional attitude.

Biconditional
°Equivalence.

Bijection Principle
SYNTAX: principle stating that an °operator must bind one and only one °variable, where 'variable' is defined as any locally A-bar bound NP. This principle (BP for short) accounts for weak °crossover violations, as in (i):

(i)  *Whoi does hisi mother admire ti
(ii)  Whoi ti likes hisi mother
The BP excludes the interpretation for (i) where both his and ti are °bound variables bound by the °operator who. The bound variable reading for the pronoun is allowed in (ii), since the pronoun, being locally A-bound by the trace of who, does not syntactically count as a variable.
LIT. Chomsky (1982), Koopman & Sportiche (1982), Van Riemsdijk & Williams (1986).

Bilabial
PHONOLOGY: bilabial sounds are produced by using both lips. EXAMPLE: in English /p/, /b/ and /m/ are bilabial.

Binary Branching Constraint
MORPHOLOGY: a constraint on concatenative °word formation which says that in the process of word formation only two morphemes can be concatenated at the same time. Hence, the compound a,b,c either has the structure [[[a] [b]] [c]], or the structure [[a] [[b] [c]]], but not the ternary structure [[a] [b] [c]]. °Circumfixes are problematic with respect to this constraint.
SYNTAX: Constraint proposed in Kayne (1984) which rules out syntactic structures in which a phrase contains more than two immediate constituents (i.e. no °node in a °tree may have more than two branches).
LIT. Spencer (1991).

Binary connective
°Connective.

Bind
SYNTAX: In °Binding Theory, a relation that may obtain between two (NP-)nodes. In Chomsky (1981), a node a is said to bind a node b iff a and b have the same °referential indexcoindexing), and a °c-commands b. Furthermore, if a binds b then, if a is in an °A-position, a A-binds b and if a is in an A-bar position, a A-bar-binds b. These notions are crucially involved in the definition of the conditions of the Binding Theory.
LIT. Chomsky (1981, 1982).

Binding
SYNTAX: A relation in which the reference of a certain element is dependent on the reference of another element. See °Binding theory, °Bind.
SEMANTICS: The relation obtaining between a °quantifier All(v) or Exists(v) and the occurrences of the °variable v in its °scope:

(i)      All(v)[ ... v ... ]
(ii)  Exists(v)[ ... v ... ]
In the following formula only the first occurrence of x is bound by All but not the second (which is not in the °scope of All):
(iii) All(x)[P(x) -> Q(y)] & R(x)
The first occurrence of x is called a bound variable, the second occurrence is called a free variable.
LIT. Gamut (1991).

Binding Condition
SYNTAX: condition on the binding possibilities of certain elements. Also called 'Binding Principle'. See °Binding theory.

Binding Domain
SYNTAX: domain in which an °anaphor or °pronoun must or must not be bound. In one version elaborated as the °Minimal Governing Category. Also called local domain. See °Binding theory.

Binding theory
SYNTAX: theory on °A-binding. Binding theory consists of the three conditions in (i).

(i)  A An °anaphor is °bound in its °binding domain
     B A °pronominal is °free in its binding domain
     C An °R-expression is free
Usually these conditions are conjoined with a characterization of anaphors, pronominals and R-expressions in terms of the features 'anaphor' and 'pronominal' as in (ii).
(ii) a [+anaphor, -pronominal]  =  anaphor
     b [+anaphor, +pronominal]  =  °PRO
     c [-anaphor, -pronominal]  =  R-expression
     d [-anaphor, +pronominal]  =  pronominal
LIT. Chomsky (1981, 1986a).

Biuniqueness
MORPHOLOGY: an axiom explicitly assumed within the framework of °Natural Morphology which entails that every morpheme has one phonological form and one meaning, and every meaning (or grammatical category) corresponds to exactly one phonological form.
LIT. Dressler (1985a, 1985b).

Bleeding order
PHONOLOGY: order of rules such that one rule destroys the input of another rule. EXAMPLE: consider the following two rules proposed by Schane (1968) for French: (a) a vowel is °nasalized before a °nasal, and (b) a nasal is dropped in syllable-final position. To derive the output [bõ] of bon, rule (a) must be applied before rule (b). If (b) would precede rule (a) it would bleed rule (a): the vowel cannot be nasalized anymore and [bõ] could not be derived.

Blending
MORPHOLOGY: an improductive type of word formation by which a new word is formed out of the initial phoneme(s) of one word and the final phoneme(s) of another. It is often argued that this type of word-formation does not belong to the I-language. EXAMPLE: A prototypical example of a blend is the English word smog which is a merger of the words smoke and fog.

Blocking
MORPHOLOGY: a process assumed in Aronoff (1976) by which the existence of one word, which is listed in the mental dictionary, prevents the application of an unproductive word formation rule, if that application would give rise to a complex word having the same semantics as the already existing word. EXAMPLE: the English suffixes -ity and -ness are very similar, the difference being that the former is unproductive and the latter productive. Both can be added to adjectives ending in -ous (curious: curiosity: curiousness). However, if there already exists a noun corresponding to the adjective in -ous as in the case of glorious: glory, gracious: grace, furious: fury etc., -ity affixation, but not -ness affixation, is disallowed: *gloriosity: gloriousness, *graciosity: graciousness, and *furiosity: furiousness.
LIT. Aronoff (1976), Spencer (1991).

Blocking Category
SYNTAX: g is a Blocking Category (BC) for b iff g is not °L-marked and g °dominates b. The concept of a BC plays a role in the definition of °barrierhood, introduced in Chomsky (1986b).

Bound
°Bind.

Bound morpheme
MORPHOLOGY: a morphological element that can only appear as a proper subpart of a word, i.e. an element which cannot function as an independent word or °free morpheme. One can distinguish two types of bound morphemes: (a) °affixes, and (b) °roots. EXAMPLE: the English word agreement contains the nominalizing affix -ment which is not a word in its own right, and therefore -ment is an affix of English. A form like mit (as in permit, remit, commit, etc.) which is not an affix but a root (since affixes may attach to it, forming a word) also cannot occur freely in syntax and may be called a bound morpheme for this reason.

Bound variable
°Variable, °binding.

Boundary
MORPHOLOGY: a formal device used in Chomsky & Halle (1968) to express a distinction between two types of affixes. The assumption that affixes are associated with different boundaries, viz. '+' (morpheme boundary) and '#' (word boundary) accounts for the fact that the English suffixes -ity and -ness behave differently with respect to a number of phonological rules, as shown in (i):

(i)	prodúctive   productívity  prodúctiveness
	op[ei]que    op[æ]city	   op[ei]queness
The claim is that -ity is a morpheme-boundary or formative-boundary affix (i.e. +ity), and -ness a word-boundary affix (i.e. #ness). °Level Ordering Hypothesis.

Bounded/unbounded
PHONOLOGY: one of the hypothesized typological parameters that define stress systems (cf. Hayes 1981)). It fixes the number of syllables of a °foot as either two (bounded) or indefinite (unbounded or n-ary).

Bounding node
SYNTAX: node that plays a role in determining whether a movement is local enough; °Bounding theory. Traditionally, NP and S (in English) or S' (in Italian) are considered bounding nodes. More recently, bounding nodes have been defined in terms of °barriers.
LIT. Chomsky (1981, 1986b), Rizzi (1982).

Bounding theory
SYNTAX: Theory about the locality of movement. The main principle of Bounding theory is the Subjacency condition, which forbids movement across more than one °bounding node. EXAMPLE: in (i) which books has been moved over two bounding nodes, NP and CP. In (ii), NP and IP are the relevant bounding nodes. In (i) the so-called °Complex NP Constraint is violated, in (ii) the so-called °Subject Condition. Thus, the Subjacency condition subsumes both the Complex NP Constraint and the Subject Condition.

(i)   *which booki did John meet [NP a child [CP who read ti]]
(ii)  *the man [CP whoi [IP [NP pictures of ti] are on the table]]]
LIT. Chomsky (1965, 1981, 1986b), Lasnik & Saito (1984, 1991), and (contra Subjacency), Bresnan (1973, 1976), Brame (1978).

Braces
PHONOLOGY: a notational convention to abbreviate two rules. EXAMPLE: braces are a means to abbreviate the rules (i) and (ii) as the one rule (iii).

(i)	A -> B	/    C   _____

(ii)	A -> B	/    D   _____

		     C 	
(iii)	A -> B	/  {   } ______
		     D 
LIT. Chomsky & Halle (1968).

Bracket Erasure Convention
PHONOLOGY/MORPHOLOGY: a convention proposed in Kiparsky (1982) stating that internal brackets are erased at the end of a lexical level or °stratum. As a consequence of this convention words become phonologically inert at the end of each lexical level, i.e., they can no longer be affected by cyclic phonological rules. After bracket erasure, morphologically derived words are treated as though they were underived. In Kiparsky's view this inertness extends to morphological processes, and word formation rules therefore do not have access to the internal structure of words derived at an earlier level. Thus, Siegel's (1977) °Adjacency Condition or Williams' (1981a) °Atom Condition can be reduced to the Bracket Erasure Convention.
LIT. Kiparsky (1982, 1985), Spencer (1991).

Bracketed grids
PHONOLOGY: a representation of stress combining aspects of grid (only) theory and °tree (only) theory. Grids as hierarchically layered structures are provided with constituent information indicated by parentheses. EXAMPLE: The stress pattern of the English word apalachicola can be represented by the following bracketed grid (cf. Halle & Vergnaud 1987):

	            *
	 *          *
	(* .)(*  .)(* .)
	 apa lachi cola
LIT. Hammond (1984), Hayes (1987)), Halle & Vergnaud (1987), Kager (1989), Hayes (1991).

Bracketing paradox
MORPHOLOGY: a situation in which the morphological structure of a word which one would like to propose for semantic reasons does not correspond to the structure one would like to propose for phonological reasons. Hence, a situation in which morphophonological structure and semantic structure are not isomorphic. The existence of bracketing paradoxes is closely associated with level-ordering theories such as Pesetsky's (1979) and Kiparsky's (1982) theories of °Lexical Phonology/Morphology. These theories propose morphological structures on the basis of level-ordering which are sometimes in conflict with the semantic structure. EXAMPLE: A well-known case of a bracketing paradox independent of °level-ordering involves the English comparative suffix -er and the negative prefix un-. In this case the phonological argumentation derives from a well-known restriction on the attachment of -er. This suffix is subject to the following constraint: it may attach to monosyllabic adjectives, and a small class of bisyllabic ones with a light final syllable, while it may not attach to adjectives with two or more heavy syllables. Compare (i):

(i)	blacker,   softer,     poorer,      nicer
	happier,   luckier,    heavier
       *directer, *complexer, *eloquenter, *importanter
Apparent counterexamples involve the prefix un-, since unhappier, unluckier etc. are well-formed. In level-ordering theories, the solution to this problem is almost trivial: assume that -er suffixation takes place at an earlier level than un- prefixation. In that case -er is added to, for instance, happy, and un- to happier, and the phonological condition on -er suffixation is not violated. Hence, the phonologically motivated morphological structure of unhappier is [un [happy-er]A]A. However, this structure raises a problem of interpretation. The meaning of the word unhappier can be paraphrased as 'more not happy', i.e. with more having °scope over un-, and crucially not as 'not more happy' with reversed scope. Hence the semantically motivated structure is [[un happy] er].
All bracketing paradoxes take the form: [prefix [root+suffix]] <-> [[prefix+root] suffix].
Another term for bracketing paradox is 'relatedness paradox'.
LIT. Pesetsky (1979, 1985), Williams (1981a), Kiparsky (1983), Sproat (1985 1988), Halle & Vergnaud (1987), Hoeksema (1985, 1987).

Branch
°Tree structure.

Bridge verb
SYNTAX: verb which allows wh-movement out of its complement. As shown by the grammaticality judgments in (i), say and think are examples of bridge verbs, while quip and whisper are not.

(i) whoi did you think/say/*quip/*whisper [ that Bill saw ti ]

By-phrase
SYNTAX: optional °adjunct in a °passive construction headed by by and containing the logical subject. EXAMPLE: in (i) the by-phrase contains Vitesse, the logical subject of won (cf. Vitesse won the match). As shown by (ii), by-phrases are also possible in nominal passives.

(i)   The match was won by Vitesse
(ii)  Carthago's destruction by the Romans
(iii) They kissed by the light of the moon
Not all languages allow passive by-phrases, and not all phrases headed by by are passive by-phrases: in (iii) the light of the moon is not interpreted as an argument of the verb kiss.
LIT. Jaeggli (1986a), Baker, Johnson & Roberts (1989), Grimshaw (1990).