We are happy to announce a talk by Ahmad Al-Bitar (GU Frankfurt) at the Semantics Colloquium.

Please register beforehand (s.walter@em.uni-frankfurt.de) to receive the access data to zoom on Thursday shortly before the talk starts.

Title: ‘Average’: A unique modifier

Date: December 17

Time: 4 pm – 6 pm ct


The modifier average has interestingly different uses, among which the ‘abstract’ use is the most well-studied. Carlson & Pelletier (2002), Kennedy & Stanley (2009) and Morzycki (2016) provide analyses of the famous example in (1-a). This sentence can be true without there being an individual with the impossible property of having 2.3 children. Hence the naming ‘abstract’.

(1) a. The average American has 2.3 children. (Abstract use)

      b. The average age of the students is 21. (Abstract use)

This use is differentiated from another use called the ‘concrete’ use as in (2). While this use is compatible with both the definite and indefinite articles, the abstract use seems to (always) prefer the definite article; see (3-a) and (3-b). This preference can be unambiguously observed in Arabic due to utilising distinct lexical items. Both mutawassiT and wasaTiyy are the equivalents of ‘abstract’ average and they only occur in definite DPs. On the other hand, ‘aadiyy’, which is an equivalent of ‘concrete’ average, can occur in any DP.

(2) An/The average German eats potato salad. (Concrete use)

(3) a. ??An/The average American has 2.3 children. (Abstract use)

      b. #An/The average age of the students is 21. (Abstract use)

Despite this (in-)definiteness distinction, most existing analyses consider the definite article vacuous in the abstract use; see Carlson & Pelletier (2002) and Kennedy & Stanley (2009). Morzycki (2016)’s analysis avoids this problem and succeeds in dealing with the definite article; nevertheless, the meaning assigned to average is not adequate enough to capture its abstract sense (Morzycki 2016, p. 13). We will see that Kennedy & Stanley (2009) provide the most appropriate analysis of abstract average in that ‘averaging’ is shown as the core of its meaning. The only missing component is, however, to integrate the definite article into the semantics of average as the abstract use seems to have a uniqueness requirement. The aim of this talk is, therefore, to end with the definiteness property being compositionally part of Kennedy & Stanley (2009)’s semantics of abstract average.