FABBS Undergraduate Research Excellence Award

Xi "Richard" Chen, Cornell University College of Human Ecology

“Explaining Recollection Without Remembering”


Context recollection refers to the conscious reinstatement of realistic details that accompanied the prior occurrence of an item. This includes objective details (e.g., position, font, and color of a word) or subjective ones (e.g., associations, thoughts, and emotions). It is used in source memory, which is knowledge about when and where an item is learned. Target recollection refers to the conscious reinstatement of presented items per se, also called item memory.

Historically, memory models have posited that accurate source memory requires accurate item memory. Dual-­recollection theory (DRT) assumes instead that source memory is primarily supported by context recollection alone while item memory uses a bivariate process of target and context recollection. These two types of recollection both produce realistic phenomenology that supports true memory for presented items. Since these are distinct processes, there are manipulations that have opposite effects on them. This gives the counterintuitive possibility toexperimentally drive the accuracy of item and source memory in opposite directions.

We tested whether DRT could explain recollection without remembering. To illustrate this effect, suppose that subjects study item lists (e.g., words such as apple, feet, goat, and house) in two distinct contexts (e.g., apple and feet are read in a male voice whereas goat and house are read in a female voice) and that they later respond to tests of item memory (e.g., old/new recognition) and source memory (e.g., recognizing items’ presentation voices). Curiously, subjects are sometimes able to remember the contexts of studied items that are not remembered on item tests (e.g., goat is incorrectly judged to be new, but its presentation voice is correctly judged to be female).

In this experiment, we compared three explanations for the effect. The first comes from DRT, which as mentioned predicts that different retrieval processes have different effects on source and item memory. In particular, target recollection supports hits and suppresses false alarms on item tests while it increases false alarms more than hits on source tests by fomenting acceptance of given sources when subjects cannot recollect contextual details.

The second explanation comes from the multivariate signal detection theory (MSDT), which maintains the item-­source memory dependency. MSDT posits that subjects apply decision criterion to the information accumulated from memory to make recognition judgments. Knowledge of particular items above the criterion leads to correct identification of the item as old or new. This model explains recollection without remembering as an instance where information for an item falls below the old/new criterion thus producing an erroneous item judgment, but falls above the source criterion thus producing a correct contextual judgment. Under this model, increasing the old/new criterion while holding the source criterion constant should also increase recollection without remembering.

The third explanation argues that recollection without remembering is merely an artifact of experimental procedures that always administer item tests before source tests. According to the artifactual explanation, the initial item test can provide subjects practice at retrieving information about an item, and hence such practice may improve memory for that item on later source tests.


Subjects studied two word lists (A and B), both accompanied by a distinct auditory context (e.g., list A was read in male voice and list B was read in a female voice). Each list was composed of words belonging to familiar taxonomic categories distinct from the other list (e.g., list A might contain multiple exemplars of fabrics and kitchen utensils presented in blocks whereas list B might contain exemplars of animals and colors).

Within each category, two types of exemplars were presented on the lists, which we call context-­consistent targets (CT) and context-­inconsistent targets (IT). For CT (e.g., pan), the exemplars appeared on the same list and in the same voice as six other blocked exemplars of its category (e.g., pan, pot, spatula, tongs, colander, cup, and stove appeared in consecutive positions on list A in a male voice). For IT (e.g., fork), the exemplars appeared on a different list and in a different voice than all other exemplars of its category (e.g., fork appeared on list B in a female voice). Thus, the correct source details for CT were the same as those for all but one other exemplars of its category. For IT, the correct source details were different than those of the other exemplars of its category.

After studying the lists, subjects responded to recognition tests for all words in counterbalanced order. This test phase consisted of 56 pairs (1 item and 1 source) of tests randomized for each subject. 25% percent of the probes were CT, 25% were IT, 25% were related distractors (RD;; unpresented exemplars of presented categories), and 25% were unrelated distractors (UD;; unpresented exemplars of unpresented categories).

We hypothesized that IT should enhance target recollection relative to CT, leading to more accurate item memory. This is because IT are list isolates that stand out against a background of exemplars of other categories, a classic target recollection phenomenon called the von Restorff effect. IT should also simultaneously impair context recollection, reducing the accuracy of source memory, because their correct source details (voice and list) are the opposite for all other exemplars of their respective categories. Note that these opposing effects of the category-­consistency manipulation on target and context recollection are memory effects rather than criterion effects.

A key DRT prediction is that recollection without remembering should be stronger for CT than for IT because lower target recollection for CT should simultaneously increase accuracy on source tests and decrease it on item tests. This prediction does not follow from the MSDT because decision criteria are normally more stringent on source tests than item tests, which would work against recollection without remembering in that explanation.

The prediction would not follow from the artifactual explanation either, as we counterbalanced the order of test questions. We decided to run a total of two experiments, identical except one manipulated question order between subjects and one manipulated it within subjects. This provided a complete approach to the artifactual explanation’s prediction that only one question order would produce recollection without remembering.


The factorial design of this experiment was 2 (word presentation order: first list vs. second list) X 2 (voice presentation order: first list vs. second list) X 2 (test presentation order: first vs. second) X 3 (item type: CT vs. IT vs. RD) X 2 (test type: item vs. source). Preliminary analyses revealed that word presentation, voice presentation, and test presentation order produced no main effects or interactions. Consequently, the main analysis for treatment effects was a 3 (item type: CT vs. IT vs. RD) X 2 (test type: item vs. source) repeated-­measures analysis of variance (ANOVA). Test presentation order did not have a main effect or interactions. Therefore our results do not support the artifactual explanation, which predicts an interaction where source accuracy for IT and CT is better when source tests are administered after item tests than when administered before.

There was a reliable main effect of item type, F(2, 76) = 50.35, MSE = .04, ηp2 = .57 and of test type, F(1, 38) = 153.21, MSE = .06, ηp2 = .80. There was also a reliable Item Type X Test Type interaction, F(2, 76) = 48.46, MSE = .03, ηp2 = .56. Post hoc analyses revealed that the ordering of mean acceptance probabilities was IT>CT>RD on item tests and CT>IT=RD on source tests. Altogether these results are congruent with DRT predictions about the category-­ consistency manipulation. According to those predictions, target recollection will be highest for IT because of the von Restorff effect. This is expected to increase item hits for IT relative to CT, whereas the context-­inconsistency is expected to lower source hits for IT relative to CT.

For direct statistical measures of recollection without remembering, we analyzed the conditional probability of a correct source judgment given an incorrect item judgment (CP). Then, we computed one-­sample t tests of CP values in order to determine if they were reliably >.5 (the guessing probability with two-­alternative source tests).

For CT, CP values (.64) were reliably >.5. Thus, for targets that had been presented on the same list and in the same voice as most of the other exemplars of their category, it was not necessary to be able to remember that an item has been presented in order to be able to remember its voice, although source memory was better when CT could be remembered as old. Furthermore, for IT the CP values (.49) were not reliably above chance. These findings do not support the traditional assumption that item memory is a necessary precondition for accurate source memory. That would lead one to expect that source memory would be better for IT than for CT, because item memory is better for IT, especially for items that were remembered as old on item tests. Instead, the opposite was true, and accurate item memory actually impaired source memory, relative to inaccurate item memory. From our data, it is difficult to see how accurate memory for a specific item’s source could be dependent on remembering that the item is old.


In our technological age, the study of memory has stretched beyond the behavioral sciences into a diversity of fields like information science, cultural studies, and neuroscience. What has entered the imagination of researchers in these disciplines is both how memory preserves the past, personal and collective, but also how memory can be distorting and misleading. The deeper implication of our research findings then is to identify an increasing need to rigorously test the eccentricities of memory and their underlying mechanisms.

In our study, the fundamental question was whether recollection without remembering is a real memory effect that requires theoretical explanation or merely an epiphenomenon. The artifactual explanation says that it is merely an epiphenomenon of test order. However, our test order manipulation never produced a main effect, and it failed to interact with any of the other design factors. In short, there was no support at all for the notion that recollection without remembering is wholly or partly a by-­product of responding to item tests before responding to source tests.

To compare DRT and the MSDT explanations, we analyzed the effect of category-­ consistency on item and source memory. Because the latter account posits that the same retrieval process underlies performance on both item and source tests, manipulations that improve item memory should also improve source memory. In contrast, DRT assumes that some manipulations can have opposite effects on target and context recollection, which drives accuracy on item and source tests in opposite directions. We saw that category-­consistency is an example of such a manipulation inasmuch as IT should exhibit enhanced target recollection but impaired context recollection relative to CT. This robust pattern was observed in both experiments.

Furthermore, our category-­consistency manipulation affected the observed levels of recollection without remembering, supporting the DRT prediction that the effect should be more robust for CT than for IT. Moreover, we found that recollection without remembering was reliable for CT but not for IT. It is difficult to reconcile those findings with the MSDT because the explanation depends on decision criterion for item tests being more stringent than for source tests. We computed the signal detection criterion statistic, C, for our data and found that the mean values for CT and IT were more stringent for item tests than for source tests, and that the differences in criterion stringency for CT versus IT items are roughly comparable. Thus, the expectations are that (a) recollection without remembering should be observed for both CT and IT because C was more stringent for item tests in both cases, and (b) the magnitude of the effect should be roughly comparable for CT and IT because the differences in C for item and source tests were roughly comparable. However, neither result was obtained.

On the whole, our data favor the view that recollection without remembering is a real memory effect, not just an epiphenomenon. Although the exact explanation of recollection without remembering remains to be determined, DRT appears to be a valuable tool for our growing understanding of memory.