I am a student of English syntax and I have a question about complement and adjunct. In this phrase

the strong influence of Latin upon English

I would think that the PP 'of Latin' is an adjunct, while the PP 'upon English' is a complement, on the ground that the phrase above can be paraphrased into "Latin has a strong influence upon English". But when I took a kind of mock test for syntax, the answer turned out that both PP are complements.

However, I have not understood why they belong to complements, and if you have another or possible answer ( if you have the reason why they both are complements), please let me understand this problem with careful analysis and explanation.

Thank you for paying attention to my question and I hope I will get an answer soon.

  • The distinction is not marked by a bright line. There are a couple of Answers here at varying degrees of technicality. My own approach to your specific case involves: 1) of X here is 'grammaticalized': it is a fixed way of expressing the source or 'Agent' of a verb-derived nominal like influence 2) upon Y is 'selected' by influence to express its goal or 'Patient'. Thus the obliques of the prepositions, Latin and English, are both in a sense 'arguments' of influence, not mere 'accidents' describing it. – StoneyB on hiatus Jul 7 '17 at 22:54
  • Adjuncts occur in clause structure ("She slept in the attic"), but your example is a noun phrase so there are no adjuncts present. What you have is a PP ("upon English") functioning within a larger PP "of Latin upon English". Both PPs are licensed by "influence" with the latter PP being a complement of "influence" and the former a complement of "Latin". – BillJ Jul 12 '17 at 7:36
  • @BillJ Sorry, but 'adjuncts' (in many functionally-oriented grammars= 'modifiers') may occur in phrases of ANY category, not just clauses (or VPs), but also NPs, APs, PPs, and even AdvPs! Semantically, 'adjuncts/modifiers' are nth-order PREDICATES that select an 'external' argument (= the phrase they 'modify'), whereas 'complements' are 'arguments' selected by their heads. As to your analysis, 'upon English' is NOT a constituent of 'of Latin upon English'. There is no PP constituent '*of Latin upon English'. Both PPs are INDEPENDENT complements (= arguments) of 'influence', as Greg Lee says. – user6814 Jul 14 '17 at 14:07
  • @BillJ Clarification: by adjuncts being, semantically, 'NTH-order predicates', I meant 'higher (than 1)-order predicates'. NO adjunct can be a FIRST-order predicate. Adjuncts are at least 2nd-order, or higher-order. First-order predicates are only predicated of SUBJECTS, generate propositions, and, typically, are VPs (except in headline-style verb-less propositional phrases like 'Trump in Paris', etc., where the PP is a 1st-order Pred). True adjuncts, on the contrary, never yield propositions when added to N's, A's, P's, etc. They only do when adjoined to a phrase that is itself propositional. – user6814 Jul 14 '17 at 14:44
  • @Sibutlasi In GP grammar, adjuncts are modifiers in the VP or clause together with related supplements. In other words, adjuncts are modifiers in clause structure. Elsewhere, we simply use the term 'modifier of'. – BillJ Jul 14 '17 at 15:35

Complements are arguments, in the sense of logic. Your example is a nominalization of the sentence "Latin influence(d) English strongly." In the sentence, "Latin" and "English" are subject and object, making them arguments and therefore complements (while "strong(ly)" is not). So, they should also be counted as complements in the nominalized form.

I think of the above reasoning as being about traditional nomenclature, not principle, because I do not know of any logical principle which can distinguish arguments from non-arguments for us.


Distinguishing between arguments and adjuncts has never been a strong point in the X-bar theory. My approach is to think about them in terms of semantic valency. Words — usually, but not necessarily, verbs — can have valences, i.e. a sort of obligatorily resolved variables. Knowing a verb implies the knowledge of its valences. For example, the verb to break has two valencies: the one who breaks and something that is broken. Once you've selected a verb, you must know who, or what, corresponds to its valencies. You cannot say Yesterday I broke without being able to specify what exactly you succeeded in breaking. Not knowing a valence of a verb makes it impossible to use the verb in a sentence correctly. Usually, semantic valencies become the subject and the direct object in a sentence (which are both complements of V, provided that you believe the Internal Subject Hypothesis), while all further information you might provide on an event goes to the adjunct area. In our example, semantically you have the right not to know at what time the breaking happened, because time is not a valence, but an adjunct, of the verb to break.

The same happens to some nouns and adjectives, especially those deriving from a verb and inheriting its valences. Back to your question: the noun influence, like the corresponding verb, has surely two valences. When you use this noun, you have to know who, or what, influenced whom (or what). Therefore, they are both complements.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.