Latent Network Analysis of Central Processing

认知过程需要时间。人类至少需要半秒才能从数字中减去“1”并输入答案，而减去“2”所需的时间要长得多。不同任务的反应时间之间的关系可以告诉我们有关底层过程组织方式的细节。这项研究将进一步发展一种特殊的方法，通过基于调度数学理论的反应时间来发现过程组织，该理论涉及如何最好地组织使用PERT（程序评估和审查技术）网络执行任务所涉及的过程。由于通常使用调度理论，人们知道在PERT网络中安排进程的方式以及每个进程需要多长时间，并且希望确定在各种条件下执行整个任务所需的时间。认知心理学家面临的是相反的问题，他们知道在各种条件下执行任务所需的时间，并想要推断网络的结构，估计每个过程所需的时间。在这一点上，我们知道这是如何在双重任务中工作的，任务中有两个刺激，每个刺激需要一个单独的反应。似乎可以同时执行多个感知过程，就像可以执行多个与运动相关的过程一样，但决定和记忆等过程必须一次执行一个。由于后一种进程需要注意，它们似乎使用一个中央处理器，一次只执行其中一个进程。一些实验将研究双重任务，重点是了解这些中心过程的执行顺序。调度理论表明，某些顺序比其他顺序更优，实验将研究受试者是自发地使用最优顺序，还是可以通过实践来学习。当然，使用最优顺序可以减少反应时间。在日常任务中，节省零点几秒或提高几个百分点的精度不会有太大的不同，但在未来的任务中，可能会有实质性的改进，比如空中交通管制，这样就有可能优化人们接收信息的类型和接收信息的顺序。其他实验将研究在阅读PASCAL计算机语言语句时心理过程的组织；在调试程序时，需要仔细阅读这些语句。最雄心勃勃的实验将研究短期记忆中回忆的过程。当一个项目列表，例如一个电话号码，被立即回忆起来时，中间的几个项目并没有被回忆起来，而是从保留在记忆中的项目的退化片段中重建出来的。通过对被试大声回忆的时间进行计时，就有可能研究这种重建过程。

Cognitive processes take time. It takes a human at least half a second to subtract "1" from a number and key in the answer, and subtracting "2" takes markedly longer. The relationships among reaction times taken by different tasks can tell us details about the way the underlying processes are organized. This research will develop further a particular method for discovering process organization through reaction times based on the mathematical theory of scheduling, a theory concerned with how best to organize the processes involved in carrying out a task using a PERT (Program Evaluation and Review Technique) network. As the theory of scheduling is ordinarily used, one knows the way in which the processes are arranged in a PERT network and how long each one takes, and one wants to determine how much time will be required to carry out the entire task under various conditions. The cognitive psychologist faces the inverse problem, knowing how long it takes to carry out the task under various conditions, and wanting to infer the structure of the network and to estimate the time required for each process. At this point, we know how this works in dual tasks, tasks in which two stimuli are presented, each requiring a separate response. It seems to be the case that more than one perceptual process may be executed at the same time, as may more than one movement-related process, but processes such as deciding and remembering must go on one at a time. Because the latter processes demand attention, they appear to use a central processor which carries out only one of them at a time. Some of the experiments will investigate dual tasks, with an emphasis on learning about the order in which these central processes are carried out. Scheduling theory indicates that some orders are more optimal than others, and the experiments will investigate whether subjects spontaneously use the optimal order, or can learn to do so with practice. Using the optimal order would, of course, reduce reaction times. Saving a few tenths of a second or improving accuracy by a few percent would not make much difference in everyday tasks, but might make a substantial future improvement in tasks such as air traffic control, in which it may be possible to optimize the type of information people receive and the order in which they receive it. Other experiments will investigate the organization of mental processes when reading statements in the PASCAL computer language; such statements need to be carefully read when debugging programs. The most ambitious experiments will investigate the processing involved in recall from short-term memory. When a list of items, a phone number for example, is recalled immediately, a few items in the middle are not so much recalled as reconstructed from the degraded fragments of the items that remain in memory. By timing subjects as the recall aloud, it is possible to study the processes involved in such reconstruction.