Historical Linguistics Compounds

Historical linguistics

Compounding in languages

A compound is a complex lexeme consisting of two or more base lexemes (compound members). Compound rules may differ in productivity: for example, in ENG the N+N rule is extremely productive, but V+N/N+V rules are unproductive and limited to a few lexical items. In other languages, N+V is possible; the process is called incorporation (the object is incorporated into the verb). In compounding, lexeme stems are combined (similar to derivational affixation): the first member is not an inflected word-form, but a stem (ex. SAN, where the first member shows a vowel-final form ‘dear to the spouse’ < ‘spouse’).

In GER this is clear in cases like Wasch-maschine or Schreib-tisch, because the verbs do not have any special suffix. In GER some compounds have a special semantically empty suffix (interfix) on the first compound member, like in Volk-s-wagen ‘people’s car’.

The first member’s function is usually to modify and narrow the denotation of the second member; the compound is a hyponym of the second member. So the second member is usually the head, and the modifying member is the dependent. In ENG the head is always the second member, but in SPA or ITA the head is the first member.

The semantic relations occurring between the members are various (purpose, appearance, location, event participant), but ultimately it is our knowledge of the world that allows us to interpret the compound correctly. Affix compounds are patterns consisting of more than one stem + an affix (ENG green-eyed, dark-haired); these are similar to AN.GR exocentric compounds, they contain the meaning element ‘having’ which can be identified in the suffix (but it’s not its only function).

Terminology

• Root: The form that remains when all derivational and inflectional elements are removed, not further analysable;
• Stem: Bound representation of lexemes that remains when the inflectional morphemes are removed;
• Words: Syntactically free simple or complex realizations of lexemes.
• Lexeme: Simple or complex dictionary entry.
• Word-forms: Inflectional forms of a lexeme.

Types of morphological stems

Depending on the morphological status, three types of morphological stems can be distinguished:

• Word-based morphology: If the base form is an entity that can function as a word in an utterance without the addition of any additional morphological material (cat-s, cheat-ed);
• Stem-based morphology: If the base form does not occur as an independent word but requires additional morphological material in order to function as a word (scient-(-ist) vs science, dramat-(-ic) vs drama);
• Root-based morphology: If the base form requires additional morphological material (a root) just to become a stem.
• Isolating language: If all bases are necessarily words and there is no affixation (Chinese).

Parallels with syntactic phrases

• Childcare care for children
• Longhouse long house

In compounds and syntactic phrases, the semantic criterion may serve to identify the head; but in syntax, there is a number of properties shared by heads:

1. The head is the morphosyntactic locus (bears inflectional markers of the whole phrase).
2. The head may govern the form of its dependents.
3. The head may agree in pers/num with its dependents.

In compounds, b. and c. cannot be observed! The dependent member is an uninflected stem whose inflectional form cannot be governed or control agreement!

A compound with more than two nouns may allow two different hierarchical structures simultaneously, like in [[nuclear power][station]] / [[nuclear][power station]]. Derivational affixes can change the word-class of their base lexeme, therefore behaving like heads of compounds.

Words or phrases?

Unclear boundaries between compounds/phrases with two content words (back bench, backdoor). Unclear boundaries between affixed words/phrases with a content word and a function word (donne-le-moi). Intermediate category of clitics, which are neither affixes nor prototypical independent word-forms; it is part of the category of bound forms (together with affixes), while independent word-forms are free forms.

Free forms vs bound forms

A free form is prosodically independent, while a bound form is prosodically dependent. Therefore, an utterance can be interrupted at a boundary between two free forms, but not at the boundary between two bound forms. In languages that use stress to express contrast, free forms can be stressed contrastively, while bound forms cannot (English); in languages that use clefting to express contrast, free forms can be clefted, but bound forms cannot (French).

Other syntactic constructions limited to free forms are topicalization and coordination. Prosodic independence also means that free forms are a separate domain for word stress, while bound forms are not. Free form boundaries are often boundaries for morpho-phonological rules that apply to combinations of free and bound forms, but do not apply across several free forms.

Clitics vs affixes

Clitics are usually defined as a type of word-forms, but they are prosodically dependent and have all the features of bound forms; so the rules governing the combination of clitics and their ‘hosts’ are in the domain of syntax, not morphology. Clitics are different than affixes in that they have freedom of movement, so they can occur in different positions in the sentence (of course only in languages which permit movement). Clitics are also less prosodically integrated with their hosts than affixes are (there are fewer prosodic rules that take the clitic group as their domain).

Compounds vs phrases

It is difficult to give a universally applicable definition of compounds: this is because the constituents are not free-standing words, but stems/roots (in some languages). Stems = uninflected parts of independent words. Compound as a whole is inflected. There are many cases in which compounds are similar to phrases with a similar meaning. The most important property of compounds is that they are idiomatic, i.e., their meaning cannot be determined from the meaning of their constituents, but it is not a sufficient criterion for compound status.

A semantic property of almost all compounds is that a dependent noun does not denote a particular referent, but an entire class (not referential but generic); in syntactic phrases, instead, a noun is typically referential. Nevertheless, we cannot say that if a dependent noun is generic then the expression is a compound, while if a dependent noun is referential, we can definitely say that the expression is a phrase (and not a compound).

Definitions of compound

• (1942) Jespersen: “A combination of two or more words so as to function as one word, as a unit”, “if the meaning of the whole cannot be deduced from the meaning of the separate elements”
• (1960) Marchand: “Two or more words combined into a morphological unit” - does not take into consideration inflectional languages (not words, but stems);
• Bauer: “A combination of two or more lexemes” the term is more appropriate, being specific enough to exclude affixes, but broad enough to include roots, stems, and free words.

Distinguishing bound roots from derivational affixes

Semantic: Roots have more semantic substance than affixes (although there are languages where this is not the case).

Formal: Roots can also occur as free forms, while affixes cannot (although in overfly and outrun the adverbs attached to the verbs behave differently than first elements of compounds usually do, aka they affect the syntactic distribution of the word *The plane flew the field The plane overflew the field).

To identify compounds we have to rely on phonological, morphological, and syntactic properties, since the typical semantic properties are not unique to compounds. In general, compounds have a greater phonological, morphological, and syntactic cohesion than phrases.

• Orthographic criterion: Separate vs one-word spelling (not useful in English); in some languages there are linking elements (GER Liebe-s-lied, the s is semantically empty and cannot be related to any inflectional form).
• Phonological criterion: Compounds are different from phrases in that they have the phonological structure of single words (usually). Stress pattern indicates compound status in languages like ENG (usually stress on the first element; however, this is not a necessary or sufficient criterion for compound status; pronunciation can vary depending on the speaker/context).

Marchand tried to identify compounds with present/past participle as the second stem as exceptions to left-hand stressing (easy-going, man-made), Giegerich tried to relate stress to the structural characteristics of N+N constructions, which are mostly phrases and therefore have stress on the right, Jones tried to identify three semantic criteria influencing stress pattern (denotation of a new idea, restriction of the meaning of the 2nd constituent).

In other languages phonological criteria are:

• Distinctive tonal patterns (Bambara, Hausa, Konni)
• Vowel harmony (Chuckchee)
• Stress patterns (German, Danish, Modern Greek)
• Segmental effects, like fricative voicing (Slave) or voicing (Japanese)
• Vowel deletion (Hebrew) or vowel reduction (Maipure, Baniva)

Morphological criterion: Compounds have greater morphological cohesion than phrases; in languages that use inflection, a compound is recognizable because its head bears the inflection for the compound as a unit and the modifying element carries some compound-internal inflection. Plural or possessive marking is possible, but not necessary in a compound (girl club already implies a plural). In verbal compounds, verbs may appear in their stem-form, infinitive, participial, or other form (rarely in finite forms). The definition of a compound as ‘a unit consisting of two or more bases’ is incorrect, because a base can also be bound! Example of pluralia tantum, words that lose their plural morpheme in composition (like oats oatmeal).

• Syntactic criterion: Compounds are inseparable (uninterruptability), while syntactic phrases are separable (a black ugly bird - an ugly blackbird); the first stem of a compound is not modifiable by adjectives or adverbs, while modification is possible in syntactic phrases (a very black bird - *a very blackbird); the second stem of a compound cannot be replaced with a proform (such as one), while this is possible in phrases (a black one - a blackone); the order of the elements is not reversible (if it is a flowerbed, it cannot be a bed flower). In addition, it is not possible for compounds to have degrees of comparison or to be