One way in which thinking about time is particularly sophisticated is when we speculate about what might have been (counterfactual thinking, e.g., If I had left earlier, I would have caught the train). Two different accounts of the development of counterfactual thinking exist. Broadly, one account argues that counterfactuals are early developing and cognitively easy, whereas the other argues that they are late developing and cognitively challenging (see a recent exchange between Beck [Reference Beck2015a; Reference Beck2015b] and Weisberg & Gopnik [Reference Weisberg and Gopnik2013; Reference Weisberg and Gopnik2015]). Hoerl & McCormack (H&M) do not discuss counterfactual thinking in their article, but the topic is worth considering, because (1) their dual systems model provides insight into the debate on the emergence of counterfactual thinking, and (2) counterfactual thinking offers additional evidence to support their approach.
If one takes the broad approach to defining counterfactuals, including all worlds that are not currently true, then many types of counterfactual thinking, such as pretence and fiction, do not require a temporal reasoning system because they have an atemporal relation to reality. These are not imagined worlds in the past or future, but they exist in parallel with the real world. They may include events unfolding in time, such as an extended pretend tea party, which can be handled using temporal updating. More relevant for this commentary is the possibility that future hypotheticals (questions about how the world will be in the future), can also be handled by the temporal updating system if one allows for imagined updates to the current world model. For example, in one study, a car was driven from the midpoint on a road to one end. Even 3-year-olds easily answered, “What if next time he drives the other way, where will he be?” by pointing to the other end of the road (Robinson & Beck Reference Robinson, Beck, Mitchell and Riggs2000; see also Beck et al. Reference Beck, Robinson, Carroll and Apperly2006). Although this appears to be speculation about a future world, it could be the output of an updated model of how the world is without any temporal reference. Reasoning with imagined extensions of actual events has been previously described as basic conditional reasoning (Perner & Rafetseder Reference Perner, Rafetseder, Hoerl, McCormack and Beck2011).
Psychologists who take the narrow approach treat tasks that involve thinking about alternatives to past events as a special type of counterfactual and assume that these thoughts about “what might have been” involve thinking about events in time (which would require the temporal reasoning system). However, 3-year-olds perform remarkably well on some of these tasks. For example, in Buchsbaum et al. (Reference Buchsbaum, Bridgers, Weisberg and Gopnik2012), children learnt that one type of block, a zando, made a machine play music. Even 3-year-olds could answer the question, “If this one were not a zando, what would happen if we put it on top of the machine?” (right answer: no music). We wonder if the question does not require thinking back in time. Despite its broadly counterfactual nature one can reach the right answer by updating a current model of the world: “Taking it from now” how will the world be if I imagine the block not being a zando?
Two recent studies have used doubly determined events, where a single outcome has more than one potential cause, to try to ensure that children had to consider the specific events in the past. In McCormack et al. (Reference McCormack, Ho, Gribben, O'Connor and Hoerl2018), two birds travelled down two separate slides towards a pig. If they reached this pig, it would get knocked over, but sometimes their paths were blocked. Children were asked various counterfactual questions, such as, “If I had not rolled the red bird that time, would the pig have fallen down?” In a study with very similar causal structure, Nyhout and Ganea (Reference Nyhout and Ganea2018) used a block-activated machine. A blue block was placed on the machine, activating it to light up, and then a green block was added (also causally effective). Children were asked, “If she had not put the green one on the box, would the light still have switched on?” Children did not give the right answer to McCormack et al.’s question until they were 6, whereas 4-year-olds gave correct answers to Nyhout and Ganea's question. Perhaps the two systems approach can help us understand this discrepancy in performance. In Nyhout and Ganea's task, the music box is presented on a video, which is stopped when children are asked the question. H&M's temporal updating system can include things outside their “current sensory scope” in the current world model. Perhaps the child's current model still includes both objects on the box so that he or she can answer the question by updating: If I remove the green block from my model, is the machine still on? A subtle difference in presentation meant that the event in the McCormack task is completed and in the past: The current model has birds at the bottom of their runs and the pig has fallen over. The child can't start from here to update her model; she needs to simulate a new version of events that are now in the past. This is beyond the temporal updating system. An important step to test this suggestion would be to contrast closely matched doubly determined scenarios where the outcome is ongoing (and can be updated) or completed (n.b. Riggs & Peterson [Reference Riggs, Peterson, Mitchell and Riggs2000] contrasted counterfactuals about ongoing and completed events, but their scenarios were singly determined and may not have necessitated thinking about events in time).
Future work on the development of counterfactual thinking will benefit from considering the temporal aspects of the demands being made on the child. We are optimistic that drawing these distinctions between different types of counterfactual thinking will be useful in understanding why some counterfactuals, those that are in and about time, are particularly challenging for children and only emerge in middle childhood when the temporal reasoning system is in place.
One way in which thinking about time is particularly sophisticated is when we speculate about what might have been (counterfactual thinking, e.g., If I had left earlier, I would have caught the train). Two different accounts of the development of counterfactual thinking exist. Broadly, one account argues that counterfactuals are early developing and cognitively easy, whereas the other argues that they are late developing and cognitively challenging (see a recent exchange between Beck [Reference Beck2015a; Reference Beck2015b] and Weisberg & Gopnik [Reference Weisberg and Gopnik2013; Reference Weisberg and Gopnik2015]). Hoerl & McCormack (H&M) do not discuss counterfactual thinking in their article, but the topic is worth considering, because (1) their dual systems model provides insight into the debate on the emergence of counterfactual thinking, and (2) counterfactual thinking offers additional evidence to support their approach.
If one takes the broad approach to defining counterfactuals, including all worlds that are not currently true, then many types of counterfactual thinking, such as pretence and fiction, do not require a temporal reasoning system because they have an atemporal relation to reality. These are not imagined worlds in the past or future, but they exist in parallel with the real world. They may include events unfolding in time, such as an extended pretend tea party, which can be handled using temporal updating. More relevant for this commentary is the possibility that future hypotheticals (questions about how the world will be in the future), can also be handled by the temporal updating system if one allows for imagined updates to the current world model. For example, in one study, a car was driven from the midpoint on a road to one end. Even 3-year-olds easily answered, “What if next time he drives the other way, where will he be?” by pointing to the other end of the road (Robinson & Beck Reference Robinson, Beck, Mitchell and Riggs2000; see also Beck et al. Reference Beck, Robinson, Carroll and Apperly2006). Although this appears to be speculation about a future world, it could be the output of an updated model of how the world is without any temporal reference. Reasoning with imagined extensions of actual events has been previously described as basic conditional reasoning (Perner & Rafetseder Reference Perner, Rafetseder, Hoerl, McCormack and Beck2011).
Psychologists who take the narrow approach treat tasks that involve thinking about alternatives to past events as a special type of counterfactual and assume that these thoughts about “what might have been” involve thinking about events in time (which would require the temporal reasoning system). However, 3-year-olds perform remarkably well on some of these tasks. For example, in Buchsbaum et al. (Reference Buchsbaum, Bridgers, Weisberg and Gopnik2012), children learnt that one type of block, a zando, made a machine play music. Even 3-year-olds could answer the question, “If this one were not a zando, what would happen if we put it on top of the machine?” (right answer: no music). We wonder if the question does not require thinking back in time. Despite its broadly counterfactual nature one can reach the right answer by updating a current model of the world: “Taking it from now” how will the world be if I imagine the block not being a zando?
Two recent studies have used doubly determined events, where a single outcome has more than one potential cause, to try to ensure that children had to consider the specific events in the past. In McCormack et al. (Reference McCormack, Ho, Gribben, O'Connor and Hoerl2018), two birds travelled down two separate slides towards a pig. If they reached this pig, it would get knocked over, but sometimes their paths were blocked. Children were asked various counterfactual questions, such as, “If I had not rolled the red bird that time, would the pig have fallen down?” In a study with very similar causal structure, Nyhout and Ganea (Reference Nyhout and Ganea2018) used a block-activated machine. A blue block was placed on the machine, activating it to light up, and then a green block was added (also causally effective). Children were asked, “If she had not put the green one on the box, would the light still have switched on?” Children did not give the right answer to McCormack et al.’s question until they were 6, whereas 4-year-olds gave correct answers to Nyhout and Ganea's question. Perhaps the two systems approach can help us understand this discrepancy in performance. In Nyhout and Ganea's task, the music box is presented on a video, which is stopped when children are asked the question. H&M's temporal updating system can include things outside their “current sensory scope” in the current world model. Perhaps the child's current model still includes both objects on the box so that he or she can answer the question by updating: If I remove the green block from my model, is the machine still on? A subtle difference in presentation meant that the event in the McCormack task is completed and in the past: The current model has birds at the bottom of their runs and the pig has fallen over. The child can't start from here to update her model; she needs to simulate a new version of events that are now in the past. This is beyond the temporal updating system. An important step to test this suggestion would be to contrast closely matched doubly determined scenarios where the outcome is ongoing (and can be updated) or completed (n.b. Riggs & Peterson [Reference Riggs, Peterson, Mitchell and Riggs2000] contrasted counterfactuals about ongoing and completed events, but their scenarios were singly determined and may not have necessitated thinking about events in time).
Future work on the development of counterfactual thinking will benefit from considering the temporal aspects of the demands being made on the child. We are optimistic that drawing these distinctions between different types of counterfactual thinking will be useful in understanding why some counterfactuals, those that are in and about time, are particularly challenging for children and only emerge in middle childhood when the temporal reasoning system is in place.