Unitization is described as the process whereby two or more previously separate items or stimulus components are integrated into a single unit, either through perceiving or conceiving of a structure (e.g., semantic meaning) that would connect the disparate units. It is generally believed that semantic association within unitized representations is stable once established. However, the opposing view has suggested that semantic association within unitized representations depends on the features of the to-be-remembered stimulus pairs, which is greatly affected by artificial factors. Thus, semantic association within the unitized representations could be broken up under certain circumstances. Nevertheless, very little experimental evidence exists to support either of the two views. More recently, it has been suggested that familiarity can be used as an indicator for whether the stimuli components are integrated into a unitized representation. It is thought that compound words are “unitized” stimuli with particular morpheme positional information. Therefore, the present study aimed to explore whether semantic unitization within compound words is stable during recognition by employing experimental manipulations involving the transposed morpheme position. Participants were instructed to remember compound words, transposed compound words, and non-compound word pairs. The familiarity-based frontal old/new effect (FN400) was obtained for compound words and transposed compound words during the 400–600 ms period, but not for non-compound word pairs. Additionally, the frontal old/new effects for compound words and transposed compound words did not significantly differ, which could underpin the facilitation of the familiarity process during associative memory of the transposed compound words. Overall, these findings suggest that semantic unitization within compound words was preserved even if the morpheme position was transposed, thus providing support for the view that semantic unitization within compound words is stable.

单元化被描述为通过感知或构想连接不同单元的结构（例如语义），将两个或多个先前分开的项目或刺激成分整合到一个单元中的过程。一般认为，统一表示内的语义关联一旦建立就稳定。然而，相反的观点认为，统一表示中的语义关联取决于待记忆刺激对的特征，而这受人为因素的影响很大。因此，在某些情况下，统一表示内的语义关联可能会被打破。然而，支持这两种观点中的任何一种的实验证据很少。最近，有人建议，熟悉度可以用作刺激成分是否被整合到统一表示中的指标。人们认为复合词是具有特定语素位置信息的“统一”刺激。因此，本研究旨在通过采用涉及转置语素位置的实验操作来探索复合词内的语义单元化在识别过程中是否稳定。参与者被要求记住复合词、转置复合词和非复合词对。在 400-600 毫秒期间，复合词和转置复合词获得了基于熟悉度的正面旧/新效应（FN400），但非复合词对没有。此外，复合词和换位复合词的正面新旧效应没有显着差异，这可能有助于促进复合词在联想记忆过程中的熟悉过程。总的来说，这些发现表明，即使语素位置被调换，复合词内的语义单元化仍被保留，从而为复合词内的语义单元化是稳定的观点提供了支持。