BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Efficiency by Construction - Fermin Moscoso del Prado Martin (Univ
 ersity of Cambridge)
DTSTART:20231124T120000Z
DTEND:20231124T130000Z
UID:TALK207091@talks.cam.ac.uk
CONTACT:Michael Schlichtkrull
DESCRIPTION:Across linguistic theories\, human language structures are rep
 resented by graphs (e.g.\, Chomsky\, 1957\, Tesnière\, 1959\, Chomsky\, 1
 995). Much research has focused on the mapping between such graphs and the
  actual sequences expressing utterances\, but less attention has been paid
  to the shapes that the graphs themselves take: their topologies. A curren
 t hypothesis argues that the structures in human language are primarily sh
 aped by language production (Bock\, 1982\, Bock & Warren\, 1985\, MacDonal
 d\, 1999\, Jäger & Rosenbach\, 2008\, MacDonald\, 2013).  Utterances are 
 planned in an incremental manner: successively incorporating chunks --eith
 er single words or larger units-- into partial syntactic structures (Bock\
 , 1982\, Bock & Warren\, 1985\, Bock\, 1987\, Levelt\, 1989\, Bock & Level
 t\, 1994\, Ferreira & Dell\,2000). Incremental construction should constra
 in the plausible probability distributions of syntactic structures. I will
  show that the topologies of actual syntactic graphs exhibit the precise d
 eviation from randomness that incremental construction predicts. This is a
  previously unknown universal regularity of human languages: Syntactic str
 uctures are constrained to a predictable topological distribution --that g
 enerated by sublinear preferential attachment (Krapivsky et al.\, 2000\, B
 arabasi & Posfai\, 2016)-- constant for all 124 languages studied\, across
  language families and modalities (spoken\, written\, and signed). It supp
 orts the hypothesis that syntactic structures are mainly shaped by languag
 e production. Furthermore\, it demonstrates how the observed efficiency of
  languages might arise without any optimization process. This finding impl
 icitly defines a data-free universal prior distribution for parse structur
 es\, with possible applications in language technologies.
LOCATION:Computer Lab\, SS03
END:VEVENT
END:VCALENDAR