Anders Holmberg @ LEL Research Seminar

Posted on February 6, 2018 by

Next Tuesday (Feb 13, 2018) LEL welcomes Anders Holmberg, Professor in Theoretical Linguistics at the Newcastle University. Holmberg’s research talk will focus on the compositionality of compound nouns in Swedish and Chinese. Here’s the abstract:


I will discuss the structure of words, particularly compound nouns, in a set of languages including Swedish and two varieties of Chinese: Mandarin and Xining Chinese. It is based on work together with Qi Wang at Newcastle University. The properties of complex words, including compound nouns, can be understood in a theory where (1), (2), and (3) are taken to be universally true:

(1) Words and phrases are formed by binary merge, forming phrases or complex words with a head.

(2) Roots have no syntactic category.

(3) A free content word consists minimally of two components, a root and a categorizer.

The grammar needs to ensure that items formed by merge are assigned a head consistently. In a phrase like [V DP] this is (arguably) not a problem: DP is maximal, V is not, so V is the head. In a compound it may be a problem: If two nouns merge, which one is the head? But if a root and a noun merge, then the root can’t be head as it has no categorial feature, so the noun is the head unambiguously. So what we expect to see is compounds consisting of a root modifier and a noun head. In some languages we do see this overtly. In Swedish a large class of nouns have the form skol-a ‘school’, flick-a ‘girl’, where the lexical part is a root and –a is a nominalizer (Josefsson 1998). In attributive compounds the modifier is a bare root, while the head is a noun: skol-flicka, ‘schoolgirl’ *skola-flicka. In Xining Chinese, common nouns are always reduplicated, unless they carry a derivational affix or are part of a compound.

(4) a. fo fo ‘spoon’ b. chei chei ‘bicycle’ c. di di ‘dish’

This can be understood if in Xining Chinese, unlike for instance Mandarin Chinese, the nominal categorizer cannot be null, and is assigned phonological features by reduplication. The reduplicant corresponds to the –a in Swedish skola. The prediction is that the modifier of an attributive compound can’t be reduplicated, but the head can, a correct prediction.

(5) a. mei hu ’ink box’ b. mei hu hu c. * mei mei hu

Other predictions made by this analysis in a theory assuming (1,2,3) will be shown to hold true as well. The theory has interesting consequences for the analysis of a class of words typical of Chinese, called bound roots or bound stems in the literature (Packard 2000); these are content words that cannot occur as free words. Another important difference between Xining Chinese and Mandarin is that Xining Chinese does not have recursive compounding. This will be discussed as well, and its consequences for compounding and word formation in general.

Josefsson, G. 1998. Minimal words in minimal syntax. John Benjamins.

Packard, J.L. 2000. The morphology of Chinese: a linguistic and cognitive approach. CUP.

The event details are as follows:

Room Booked: Sam Alex_A101
Date: Tue 13 Feb 2018
Time: 16.15 – 17.30

As usual, the talk will be concluded with a wine reception. We look forward to meeting you there!