The problem of serial order in behavior: Lashley’s legacy
Introduction
In 1951, Karl Lashley, a neurophysiologist at Harvard University, published a paper that has become a classic: “The Problem of Serial Order in Behavior.” The paper challenged the prevailing view of behavioral sequencing in America in the first half of the 20th Century. According to that view, expressed by Washburn, 1916, Watson, 1920, and other behaviorists, the functional mechanism that governs the unfolding of behavior is the reflex chain: Stimulation caused by movement n triggers movement n + 1, stimulation caused by movement n + 1 triggers movement n + 2, and so on.
Lashley argued against this view on the basis of three main lines of evidence: (1) Movements can occur even when sensory feedback is interrupted. (2) Some movement sequences occur too quickly for elements of the sequences to be triggered by feedback from the preceding elements. (3) Errors in behavior suggest internal plans for what will be done later.
To these sources of evidence, new sources have since been added: (4) The time to initiate a movement sequence can increase with the length or complexity of the sequence (e.g., Henry and Rogers, 1960, Klapp, 1977, Rosenbaum, 1987, Sternberg et al., 1978). (5) The properties of movements occurring early in a sequence can anticipate later features (e.g., Cohen and Rosenbaum, 2004, Kent, 1983, Van der Wel and Rosenbaum, 2007). (6) Neural activity can indicate preparation of upcoming behavioral events, including upcoming behavioral events in the relatively long-term future (e.g., Aldridge et al., 1993, Ashe et al., 2006, Averbeck et al., 2005, Kennerley et al., 2004).
All of these findings substantiate Lashley’s (1951) hypothesis that the solution to the serial order problem (i.e., the solution to the problem of how behaviors are sequenced) need not be ascribed to triggers from sensory feedback. Instead, the state of the nervous system can predispose the actor to behave in particular ways in the future. The common-sense way of saying that the nervous system prepares some behaviors but not others is to say that there are plans for behavior. Asserting that there are plans need not imply that sensory feedback plays no role, or only a minimal role, in movement control. It is known that neural states permitting differentiation of behavior can do so on the basis of sensory feedback. For example, whether a cat extends or flexes its limb when pressure is applied to its paw depends on where the limb is in the step cycle (Forssberg, Grillner, & Rossignol, 1975).
Lashley (1951) was instrumental in promoting the notion of hierarchical organization of plans. Several lines of evidence support this notion. One is that the same behaviors can have different functional interpretations depending on the context in which they occur. Thus, the sound pattern/rajt/ can be interpreted differently depending on where it occurs in a sentence, as in this sentence used by Lashley: “The mill-wright on my right thinks it right that some conventional rite should symbolize the right of every man to write as he pleases” [p. 116]. Such contextual dependence, Lashley argued, is only possible with functionally overarching states of the sort implied by hierarchical plans.
Other sources of evidence for hierarchical plans have since been added to Lashley’s list. One pertains to errors in behavior. Analyses of such errors, whether in the domains of speech (Dell, 1986, Garrett, 1975), typewriting (see Rosenbaum, 1991, Chapter 8), or everyday action (Norman, 1981), implicate hierarchical organization. To cite one well known example from language production, word exchanges rarely cross syntactic class boundaries: Nouns switch with nouns and verbs switch with verbs, but nouns and verbs rarely switch with each other (Garrett, 1975). In addition, phonological errors indicate dominance relations. If a sports announcer accidentally says that there are “two out and one runs” while narrating a baseball game, the “s” in “runs” is pronounced like a “z”, not like a hard “s,” as befits the plural form of “out” in English. This outcome suggests that phonological rules, or their neural equivalents, are invoked during speech.
A second source of evidence for hierarchical organization of plans comes from the timing of behavioral sequences. As mentioned above, the initiation time of a movement sequence can increase with its length. So too can the inter-response times of the sequence elements (Sternberg et al., 1978; for a recent review see Rhodes, Bullock, Verwey, Averbeck, & Page, 2004). Further, inter-response times can depend on the size of the phrase that is about to be generated. The larger the phrase, the longer the inter-response time (Collard and Povel, 1982, Rosenbaum et al., 1983). Such data have been interpreted in terms of ‘decoding’ or ‘unpacking’ hierarchical plans into their constituents.
A third source of evidence for the hierarchical organization of plans comes from the relative ease or difficulty of learning different kinds of behavioral sequences. It is easy to learn sequences such as 123, 234, 345, 456, …, but it is hard to learn random sequences such as 143, 256, 265, … (Restle, 1970, Simon, 1972). Sequences of the former kind are well described by rules, and may be compactly represented by computer programs, such as the following program written in MATLAB (see Rosenbaum, 2007),
for i = 1:5
for j = 1:3
y(i, j) = i + (j − 1);
end
end
y
The output from this program is
y =
1 2 3
2 3 4
3 4 5
4 5 6
5 6 7
Saying that a plan may be likened to a computer program need not imply that there is a little person in the head (a homunculus) ‘writing the program,’ although some critics of the cognitive approach to motor control have claimed that this is the case. If one conceives of a plan or ‘motor program’ as a memory for what is to come, the concept of a plan or motor program need be no more contentious than the concept of a memory for what has happened. The latter concept is not one that anyone doubts. Recent evidence linking the hippocampus, a brain area associated with memory, with the ability to imagine the future supports this analogy (Hassabis, Kumaran, Vann, & Maguire, 2007).
A fourth and final source of evidence for hierarchical plans pertains to the fact that the long-term learning of skills is naturally characterized by the process of forming ever larger hierarchical units or ‘chunks’ (Miller, 1956; for a review see Rhodes et al., 2004). According to this view, the way learners acquire new skills suggests that they learn control structures for successively larger units of behavior, with newly learned routines calling up or relying on more elementary routines (Bryan & Harter, 1897). This account has been used to interpret the observation that the growth of skill is expressed by the capacity to display increasingly complex behaviors. Thus, one learns how to play a single note on the piano and then a simple series of notes before learning to play Rachmaninov’s Fourth Piano Concerto.
The chunking perspective has also been used to account for the fact that the time to repeatedly perform a task decreases as the task is practiced more and more, with the rate of reduction decreasing as practice continues. This phenomenon is quantitatively described by the Power Law of Learning (Crossman, 1959), a law so named because the time to perform a repeatedly practiced task can be expressed mathematically as the number of times the task is practiced, raised to a numerical power. The smaller the power (i.e., the more negative the power), the greater the speeding with practice. The success of the Power Law of Learning – indeed, it is being called a Law – has been ascribed to mastering ever more inclusive routines for behavioral control (Newell & Rosenbloom, 1981).
With all this support for Lashley’s position, what more has our field contributed to Lashley’s proposal? In this paper, we present some new sources of evidence for hierarchical organization of motor plans, including results that suggest that even single movements rely on plans which are hierarchically organized. We particularly focus on investigations that have helped shed light on the processes by which plans are assembled. We have been led to the view that motor plans are assembled by gradually changing the control parameters that distinguish what needs to be done from what has just been done before. This view has come from work on sequential effects in performance, which we review in the next major section of this paper. Then we turn to the evidence suggesting that plans for single movements are hierarchically organized. This evidence comes from computational modeling and from a variety of behavioral and neurophysiological findings, to which we turn before the Conclusions section.
Section snippets
Sequential effects
One way to study motor planning is to study sequential effects in performance. We have found this to be a rich mine for investigating the processes underlying motor planning.
Motor equivalence in handwriting
The foregoing conclusions not only reveal how successive movements are generated. They also have implications for our understanding of the capacity to achieve the same physical tasks in essentially the same ways when using different effectors. Such ‘motor equivalence’ is often discussed in the context of handwriting. People can write their names or other scripts in ways that preserve their individual writing styles when they write with different muscle groups – for example, when using the
End-state comfort
In the previous section, we said that Donders’ Law as applied to arm positions is strongly violated. In this section, we review the evidence to which we referred (also see Soechting et al., 1995). The observations to be reviewed were important in the original postulation of the posture-based motion planning theory (Rosenbaum, Engelbrecht, Bushe, & Loukopoulos, 1993), and for present purposes are important for joining the discussion of the serial ordering of complex behavioral sequences (the
Conclusions
Karl Lashley (1951) challenged researchers of his day to give up the view of behavioral sequencing that prevailed at the time (reflex chaining). He also implicitly challenged researchers of succeeding generations to fill in the gaps in the understanding of the serial ordering of behavior. In the present article, we have reported our and others’ attempts to take up that challenge. We have done so by briefly reviewing findings that have supported Lashley’s hypothesis that there are hierarchically
References (85)
- et al.
Cortical control of motor sequences
Current Opinion in Neurobiology
(2006) - et al.
A vector-integration-to-endpoint model for performance of viapoint movements
Neural Networks
(1999) - et al.
Complex movements evoked by microstimulation of precentral cortex
Neuron
(2002) - et al.
The codability of kinesthetic location and distance information
Acta Psychologica
(1972) Task switching
Trends in Cognitive Sciences
(2003)- et al.
Learning and production of movement sequences: Behavioral, neurophysiological, and modeling perspectives
Human Movement Science
(2004) - et al.
From cognition to biomechanics and back: The end-state comfort effect and the middle-is-faster effect
Acta Psychologica
(1996) - et al.
The parameter remapping effect in human performance: Evidence from tongue twisters and finger fumblers
Journal of Memory and Language
(1986) Memory for movements
- et al.
On the continuity of mind: A dynamical account of cognition
The Psychology of Learning and Motivation
(2004)
The latency and duration of rapid movement sequences: Comparisons of speech and typewriting
The simultaneous chain: A new approach to serial learning
Trends in Cognitive Science
Neuronal coding of serial order: Syntax of grooming in the neostriatum
Psychological Science
Kinematic features of unrestrained arm movements
The Journal of Neuroscience
Activity in prefrontal cortex during dynamic selection of action sequences
Nature Neuroscience
Working memory
Studies in the physiology and psychology of the telegraphic language
Psychological Review
Evidence of motor planning in infant reaching behavior
Psychological Science
Where objects are grasped reveals how grasps are planned: Generation and recall of motor plans
Experimental Brain Research
Theory of serial pattern production: Tree traversals
Psychological Review
A theory of the acquisition of speed skill
Ergonomics
A spreading activation theory of retrieval in sentence production
Psychological Review
Testing hypotheses and the advancement of science: Recent attempts to falsify the equilibrium point hypothesis
Experimental Brain Research
Phase dependent reflex reversal during walking in chronic spinal cats
Brain Research
Spatial conceptual influences on the coordination of bimanual actions: When a dual task becomes a single task
Journal of Motor Behavior
The analysis of sentence production
Principles for the control of kinematically redundant limbs
Grounding language in action
Psychonomic Bulletin & Review
Signal-dependent noise determines motor planning
Nature
Patients with hippocampal amnesia cannot imagine new experiences
Proceedings of the National Academy of Sciences
Increased response latency for complicated movements and a “memory drum” theory of neuromotor reaction
Research Quarterly
An organizing principle for a class of voluntary movements
The Journal of Neuroscience
The Theory of Event Coding (TEC): A framework for perception and action planning
Behavioral And Brain Sciences
Eye movements and eye-hand coordination during typing
Current Directions in Psychological Science
Hand path priming in manual obstacle avoidance: Evidence that the dorsal stream does not only control visually guided actions in real time
Journal of Experimental Psychology: Human Perception and Performance
Computational motor control and human factors: Modeling movements in real and possible environments
Human Factors
Intersegmental coordination during reaching at natural objects
The timing of natural prehension movement
Journal of Motor Behavior
Organization of action sequences and the role of the pre-SMA
Journal of Neurophysiology
The segmental organization of speech
Reaction time analysis of programmed control
Exercise and Sport Sciences Reviews
Cited by (203)
Autonomous behaviour and the limits of human volition
2024, CognitionThink then act, or act then think? Double-response reaction times shed light on decision dynamics in precrastination
2024, Cognitive Control of Action: Selected Works of David A. RosenbaumRhythm and the embodied aesthetics of infant-caregiver dialogue: insights from phenomenology
2024, Phenomenology and the Cognitive SciencesPreliminary Insight on Neural Correlates of Quranic Impacts on Cognition: A Review
2023, Malaysian Journal of Medicine and Health Sciences