Lay Definitions of Intelligence, Knowledge, and Memory: Inter- and Independence of Constructs

Abstract

The present study examined how lay participants define the following concepts used widely in psychology: being intelligent, knowing, and remembering. In the scientific community, knowledge overlaps with the contents of semantic memory, crystallized intelligence reflects the accumulation of knowledge, knowledge and event memory interact, and fluid intelligence and working memory correlate. Naturally, the lay public has implicit theories of these constructs. These theories mainly distinguish between intelligent and unintelligent behaviors and tend to include characteristics outside psychometric studies of intelligence, such as emotional intelligence. Here, we asked lay participants from the online platform Prolific to explain “what does being intelligent mean to you?” as well as “knowing” and “remembering” to understand their degree of alignment with theoretical conceptualizations in the research community. Qualitative coding of participant definitions showed that intelligence and knowledge are closely related, but asymmetrically—when defining what it means to be intelligent, participants reference knowledge, but intelligence is not considered in explaining knowing. Although participants note that intelligence is multi-faceted and related to problem-solving, there is an emphasis (in terms of frequency of mentions) on the crystallized side of intelligence (i.e., knowledge). A deeper understanding of lay participants’ mental models of these constructs (i.e., their metacognitions) is essential for bridging gaps between experts and the general public.

Mapping the Mind With Free Associations: A Tutorial Using the R Package associatoR

“…The six clusters displayed in Figure 3B help to understand what GPT-4 (or a comparable sample of humans) sees as relevant components of the concept of intelligence. Whereas past work using manual ratings of open-ended questions identified knowledge, which is similar to our cluster Learning, and understanding as key aspects of people's mental representation of intelligence (Coane et al, 2023), the clusters highlight four further topics: innovation, research, reason-ing, and cleverness. This demonstrates the power of free associations to reveal the breadth of mental representations.…”

“…In our synthetic example, we adopted the same minisnowball design as , and focused on another central construct in cognitive psychology: intelligence. Investigating the mental models underlying this construct (Coane et al, 2023;Sternberg et al, 1981) can help to bridge gaps between expert and public understandings of the construct, and facilitate comparisons between human and machine intelligence (Cave & Dihal, 2019). For our demonstration purposes, we generated 1,000 synthetic participants using a large language model, GPT-4-Turbo (gpt-4-0125-preview;OpenAI, 2023), and used prompting to introduce group variation, creating four demographic groupings that varied in gender (male versus female) and level of education (high school versus university).…”