From stork to baby: Semantic relatedness can improve order memory without grouping

The literature suggests that semantic relatedness such as semantic association has contrasting effects on working memory (WM): It enhances item memory while it impairs or does not affect order memory. The detrimental or null effect on order memory has been widely accepted and regarded as a benchmark finding. Therefore, it is theoretically important to test the opposite effect, that is, the possibility that semantic relatedness improves order memory. As a form of semantic relatedness, the present study focuses on the directionality of associations. For example, "stork" is associated with "baby" in a forward direction (stork -> baby), but not in a backward direction (stork <- baby). Critically, we hypothesized that associative directions, as a constraint on order, would improve short-term order information recall. Four experiments used six-word lists in which all words had forward associations such as "governance -> ruling -> judge -> robe -> shower -> hot" for immediate serial recall. All experiments converged to show that directional associations indeed improved order memory. Specifically, lists with forward associations led to better order memory, relative to those with non-associations (Experiments 1-3) and those with backward associations (Experiment 4). The results sharply contrast with a widely accepted view that semantic relatedness is neutral or detrimental for order memory. Our novel findings therefore challenge existing models based on the null or detrimental effect. We also discuss how models can potentially explain our findings by incorporating representations of associative directions and/or adopting the process of associative chaining.