Project Description

One of the most intriguing features of languages is that speakers can always produce novel grammatical utterances that they have never heard before. Consequently, most linguists agree that the mental grammars of speakers are complex systems that must be more abstract than the input they are exposed to. Yet, linguists differ as to how general and abstract speakers’ mental representations have to be to allow this grammatical creativity. In order to shed light on these questions, the present project looks at one specific construction type, English comparative correlatives. A simple example of a comparative correlative construction is the more you eat, the fatter you'll get.

FORM and MEANING

[the [more]_{comparative phrase1} you eat,]_C1 [the [fatter]_{comparative phrase2} you'll get]_C2

• two clauses: C1 (the more you eat) and C2 (the fatter you'll get)
• C2 refers to a dependent variable (the fatter you'll get), which is affected by an independent variable (the more you eat).

Structural Variation: selected properties

• Open "slots"  in the construction can be filled variably:
  • the [older]_AdjP the man got,]_C1 [the [longer]_AdvP he slept.]_C2
    (Adjective Phrase and Adverb Phrase)
  • [the [more money]_NP we come across]_C1 [the [more problems]_NP we see]_C2
    (Noun Phrase and Noun Phrase)
  • [The [more under the weather]_PP you are,]_C1 [the [more in pain]_PP you are]_C2
    (Prepositional Phrase and Prepositional Phrase)

   

• Parts of the construction can be 'deleted' and even truncated:
  • main verb BE can be suppressed/deleted in either clause:
    [The higher the price is,]_C2 [the more interesting the product is.]_C2 
    [The higher the price   ,]_C2 [the more interesting the product is.]_C2
    [The higher the price is,]_C2 [the more interesting the product   .]_C2
    [The higher the price   ,]_C2 [the more interesting the product   .]_C2
  • each clause can be truncated down to its initial comparative phrase:
    [The higher the price is,]_C2 [the more interesting                      ]_C2
    [The higher                 ,]_C2 [the more interesting the product is.]_C2
    [The higher                 ,]_C2 [the more interesting                     .]_C2

   

Objectives

Previous pilot studies on CC constructions (Hoffmann 2013, 2014a,b, 2015) were based on fairly small data bases (classical one-million word corpora, which only contained 30-40 CC tokens each). In the present project, we will now draw on large samples of authentic corpus data (the British National Corpus (BNC), the Corpus of Contemporary American English (COCA), and the Global Web-Based English (GloWbE)) as well as Magnitude Estimation acceptability experiments. Following Hoffmann (2014a,b), the present study will analyse the Comparative Correlative construction using a combination of Usage-based Construction Grammar and New Englishes approaches. This approach will contribute to answering the following questions about the abstract mental representations that have to be postulated by a Construction Grammar approach as well as about the evolution of New Englishes:

1. How can we best describe the Comparative Correlative construction's abstract representation (and abstract representations in general)?
2. Are these representations the same across varieties, or do speakers of different varieties possess different representations?
3. Does the evolutionary state of the New English varieties correlate with certain types of Comparative Correlative constructions?

Project Phases

Duration of project

2017-2020 funded by the German Research Foundation (DFG)

The project has the following two key stages:

Stage I: Corpus study

• The following corpora will be queried for comparative correlatives:
  • British National Corpus (BNC)
  • Corpus of Contemporary American English (COCA)
  • Global Web-Based English (GloWbE)
• All relevant data will be annotated for a wide range of factors, including VARIETY, FILLER TYPE C1 (AdjP, AdvP , etc.), DELETION C1 (full clause (without auxiliary), be-retained, be-deletion, truncated) etc.
• The coded data will be subjected to multifactorial statistical analyses.

Stage II: Experimental study

• Speakers’ judgments on structures will be elicited using Magnitude Estimation acceptability experiments.
• Test subjects will be recruited from the following six partner universities:
  the University of Edinburgh (UK), the University of Texas at Austin (USA), the National University of Singapore (SING), the University of Hong Kong (HK), the University of Cape Town (SA) and the University of Nairobi (KE).