ChatGPT can search a large amount of data and, on the basis of the statistical probability of possible collocations, can put together something that seems like human-produced text, which should not be surprising, as the data used is human text. But these algorithms do not understand what they are writing, and cannot go beyond Artificial Narrow Intelligence. To achieve Artificial General Intelligence will require algorithms to imitate human meaning creation, which depends on abductive inference of the communicative intention of the communicator by the addressee. Abductive inference is essentially guesses as to why the communicator did what they did (which may be languaging or just gesture or facial expression, or all three). It is causal reasoning, asking why some observed phenomenon is the way it is, and this is applied in creating general knowledge, and in inferring the motivations of others when they perform some action, including its use in communication. This causal reasoning is the core of our consciousness.

 

 

Randy J. LaPolla (PhD in Linguistics, University of California, Berkeley) is a Professor in the Center for Language Sciences, Institute for Advanced Studies in Humanities and Social Sciences, Beijing Normal University, Zhuhai Campus, China, a Cheung Kong Scholar, and a Fellow of the Australian Academy of the Humanities. His research started with a focus on the history and typology of Sino-Tibetan and Austronesian languages but expanded to issues related to the nature of communicative behavior and functional explanations for the patterns found in human behavior generally and the cognition that motivates it.

 

Participants may register through the link go.ateneo.edu/Lecture-ArtificialNarrow or by scanning the QR code on the poster.