Lexical Insertion (2)

Linguistics 322


At one time in the history of transformational syntax, it was believed that lexical insertion apply simultaneously. This view has been challenged on various fronts. We will challenge this view in what follows. We took this view in our initial discussion on lexical.insertion.

We hold the view here that lexical insertion applies in certain cases after movement and insertion have taken place. The reason for our view is found basically in English morphology. Consider the following example containing the irregular verb hold. When tense is lowered to V (see ), it is adjoined to V. The first problem is what is the form of the past tense? It has various forms including '-ed', '-t', '-d', or no affix (some linguists consider such verbs to have a null affix). Usually the default form '-ed' was selected; e.g.: hold+ed. However, *holded does not occur in English. Somehow, 'hold+ed' had to be changed to 'hel+d'. Because there is no regular phonological rule which converts 'hold+ed' into 'hel+d', it was assumed that a 'second lexical pass' was necessary.

For several decades it has been argued that inflectional endings are made up of a bundle of grammatical features. The past tense is composed of one feature: [+Past]. It is uncertain at this time whether the fact that the host for [+Past] must be a verb is universal or whether it is a property of the English tense system. Once [+Past] is adjoined to the verb, how it is spelled out depends on a lexical property of the verb (see the lexicon). The default is the ending '-ed'. This property is left unmarked. If the lexical entry for the verb carries no information on the selection of the past tense suffix, then the default ending is selected. The verb hold selects the ending '-d'.

(1)    hold+[+Past] --> hold+d --> held+d --> hel+d.

The larger problem is two-fold: how to represent hold and how to account for change in the root vowel (sometimes called 'ablaut'). We will defer the first question and represent it as 'HOLD'. 'HOLD' can be thought of as an abbreviation for the features which constitute the verb hold. HOLD has two forms in English: hold and held. HOLD is spelled out as held (the orthographic form of /hEld/ if it precedes the suffix 'd' (/E/ represents a lax 'e'). Otherwise it is spelled out as hold (/hold/):

(2)    HOLD + [+Past] --> HOLD + d --> held+d.

(3)    HOLD --> hold (elsewhere).

Obviously, these two rules are ordered. 'Elsewhere' rules always apply last. Note that the final 'd' in the stem is deleted. It is deleted by a general rules that simplifies geminate obstruents in the same syllable. We need not include this in the lexical entry since it is a general and predictable rule:

  1. C --> '0' (null) / ___ C $
where both Cs represent the same obstruent, and $ represents a syllable boundary.

If the lexicon is fully specified as mentioned in the discussion on the lexicon, all these rules can be avoided.Let us start this discussion with an intransitive verb:

(4)     John swam.

The conceptual form is:


We cannot go directly to the lexicon and selected that appropriate form for the past tense ending. There are thraee endings: '-ed', '-d', '-t'. '-ed' is the default. We do not know which one to chose, since tghe ending is determined by the verb. As long as '-ed' is not adjoined to a stem, there is no way to determine this. The one thing that all these forms share is the feature [affix], which means that the form needs a host. Let us go one step further and claim that the underlying feature is [+Weak]. Weak forms require a host.x

Although there are many words in English that have no irregular forms, we will assume that all spell out rules apply after movement.

To go to ambiguity click here.

To return to

222 course outline click here

322 course outline click here