Date of Award

Spring 1-1-2015

Document Type


Degree Name

Master of Arts (MA)



First Advisor

Randy O'Reilly

Second Advisor

Matt Jones

Third Advisor

Tor Wager


Chunking is widely theorized to be a fundamental building block of cognition, underlying language, motor control and general problem solving abilities. Recent studies of chunking have focused on a sequence learning paradigm known as the Discrete Sequence Production task, which is thought to reliably elicit chunks. These chunks have been documented in sequences that are executed at remarkable speeds, greatly exceeding the documented abilities of working memory in simple reaction time task contexts. These chunks further greatly exceed working memory capacity, with previously documented chunk sizes consisting of up to ten elements. Here, we question whether what other researchers have identified as chunks really are chunks, demonstrate that subjects can create putative chunks on demand, and further demonstrate that subjects may be able to create chunks of up to size 20 with extensive practice, with the salient prediction that there is no limit to chunk size. These findings question theories of chunking that rely on working memory, question whether chunk detection algorithms are tagging random features as chunks, and point to directions for future research that might clarify these issues.


Author also known as B. J. Mingus.

Available for download on Sunday, August 22, 2117