Overview[edit]
Procedural knowledge (i.e., knowledge-how) is different from descriptive knowledge (i.e., knowledge-that) in that it can be directly applied to a task.[2][4] For instance, the procedural knowledge one uses to solve problems differs from the declarative knowledge one possesses about problem solving because this knowledge is formed by doing.[5]
The distinction between knowing-how and knowing-that was brought to prominence in epistemology by Gilbert Ryle who used it in his book The Concept of Mind.[3]
Know-how is also often referred to as street smarts (sometimes conceived as the opposite of book smarts), and a person employing their street smarts as street wise. Know-how is often tacit knowledge, which means that it can be difficult to transfer to another person by means of writing it down or verbalising it. The opposite of tacit knowledge is explicit knowledge.
Definition[edit]
Procedural knowledge is the "know how" attributed to technology defined by cognitive psychologists, which is simply "know how to do it" knowledge. Part of the complexity of it comes in trying to link it to terms such as process, problem solving, strategic thinking and the like, which in turn requires distinguishing different levels of procedure.[6] It is the ability to execute action sequences to solve problems. This type of knowledge is tied to specific problem types and therefore is not widely generalizable.[7] Procedural knowledge is goal-oriented and mediates problem-solving behavior.[8]
The concept of procedural knowledge is also widely used in mathematics educational researches. The well-influential definition of procedural knowledge in this domain comes from the introductory chapter by Hiebert and Lefevre (1986) of the seminal book "Conceptual and procedural knowledge: The case of mathematics", in which they divided procedural knowledge into two categories. The first one is a familiarity with the individual symbols of the system and with the syntactic conventions for acceptable configurations of symbols. The second one consists of rules or procedures of solving mathematical problems. In other words, they define procedural knowledge as knowledge of the syntax, steps conventions and rules for manipulating symbols.[9] Many of the procedures that students possess probably are chains of prescriptions for manipulating symbols. In their definition, procedural knowledge includes algorithms, which means if one executes the procedural steps in a predetermined order and without errors, one is guaranteed to get the solutions, but not includes heuristics, which are abstract, sophisticated and deep procedures knowledge that are tremendously powerful assets in problem solving. [10] Therefore, Star (2005) proposed a reconceptualization of procedural knowledge, suggesting that it can be either superficial, like ones mentioned in Hiebert and Lefevre (1986), or deep.[11][9] Deep procedural knowledge is associated with comprehension, flexibility and critical judgement. For example, the goals and subgoals of steps, the environment or type of situation for certain procedure, and the constraints imposed upon the procedure by the environment.[12] Research on procedural flexibility development indicates flexibility as an indicator for deep procedural knowledge. Individuals with superficial procedural knowledge can only use standard technique, which might lead to low efficiency solutions and probably inability to solve novel questions. However, more flexible solvers, with a deep procedural knowledge, can navigate their way through domain, using techniques other than ones that are over-practiced, and find the best match solutions for different conditions and goals. [13][11][14]
Development[edit]
The development of procedural knowledge is always entangled with the development of declarative knowledge. Researchers suggested that initial problem solving involves explicitly referring to examples, participants start with pure example-based processing.[15][16] The examples illustrate the solution of a similar problem and the problem solver analogically maps the solution of the example onto a solution for the current problem. People make extensive reference to examples even when they are initially taught the rules and principles.[17] It is believed that when people acquiring cognitive skills, firstly an example is encoded as a declarative structure. When participants are tested on their first problems, they have two possible ways to respond. If the example matches the problem they learned, they can simply retrieve the answer. However, if it does not match, they must analogically extend the example.[16] With repeated practice, general rules develop, and the specific example is no longer accessed. In this way, knowledge transitions from a declarative form (encoding of examples) to a procedural form (productions rules), which is called the adaptive control of thought—rational (ACT-R) theory.[18]
However, in certain occasions, procedural and declarative knowledge can be acquired independently. Researches with amnesic patients found that they can learn motor skills without the ability to recollect the episodes in which they learned them, also learned and retained the ability to read mirror-reversed words efficiently, yet were severely impaired in recognizing those words, which give evidences about the neurological basis differences in procedural and declarative knowledge.[19][20] Researchers also found that some normal subjects, like amnesic patients, showed substantial procedural learning in the absence of explicit declarative knowledge. Even though declarative knowledge may influence performance on a procedural task, procedural and declarative knowledge may be acquired separately, one does not need to have knowledge of one type in order to build the other type of knowledge. The influence caused by declarative knowledge might be due to the facilitation to a process of pathway activation that is outside of conscious awareness.[21] If the prime is highly predictive of the target, the amount of facilitation is increased because of an active, conscious, attentional effect that is superimposed on the pathway activation.[22] Therefore, if and when subjects develop explicit declarative knowledge of procedure, they can use this knowledge to form attentional expectancies regarding the next item in this procedure.[21]
Activation[edit]
Lashley (1951) proposed that behavioral sequences are typically controlled with central plans, and the structure of the plans is hierarchical. Some evidences also support this hypothesis. Same behaviors can have different functional interpretations depending on the context in which they occur. The same sound pattern can be interpreted differently depending on where it occurs in a sentence, for example, there and their. Such contextual dependence is only possible with functionally overarching states of the sort implied by hierarchical plans. [23] The initiation time of a movement sequence and the inter-response times of the sequence elements can increase with its length.[24] 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.[25][26] Such data have been interpreted in terms of decoding or unpacking hierarchical plans into their constituents. Moreover, learning difficulties changes with the easiness of behavioral sequences.[27][28] Finally, long-term learning of skills is naturally characterized by the process of forming ever larger hierarchical units or chunks.[24] People learn control structures for successively larger units of behavior, with newly learned routines calling up or relying on more elementary routines, like learning to play simple notes before being able to play a piano concerto.[29]
As for process of behavior plan forming, Rosenhaum et al. (2007) proposed that plans are not formed from scratch for each successive movement sequence but instead are formed by making whatever changes are needed to distinguish the movement sequence to be performed next from the movement sequence that has just been performed.[30] There are evidences found that motor planning occurs by changing features of successively needed motor plans.[31] Also, Rosenhaum et al. (2007) found that even single movements appear to be controlled with hierarchically organized plans, with starting and goal postures at the top level and intermediate states comprising the transition from the starting to the goal at the lower level.[30]
Interaction with conceptual knowledge[edit]
The most common understanding in relation to the procedural and conceptual knowledge is of the contrast of knowing how and knowing that.[32] Some see the distinction as a contrast between the tacit knowledge of technology and the explicit knowledge of science.[33] Conceptual knowledge allows us to explain why, hence the distinction of "know how" and "know why".[34] Conceptual knowledge is concerned with relationships among items of knowledge, such that when students can identify these links, it means they have conceptual understanding. Cognitive psychologists also use the term declarative knowledge to contrast it with procedural knowledge, and define it as "knowledge of facts".[35] However, declarative knowledge may be a collection of unrelated facts, whereas conceptual knowledge puts the focus on relationships.[36] Also, declarative knowledge is an inert form of knowledge which contrasted with procedural knowledge as an active form, but conceptual knowledge can be part of an active process. Therefore, it is important to know that conceptual knowledge is not simply factual knowledge but consists of ideas that give some power to thinking about technological activity.
Evidence from mathematics learning research supports the idea that conceptual understanding plays a role in generation and adoption of procedures. Children with greater conceptual understanding tend to have greater procedural skill.[37] Conceptual understanding precedes procedural skill.[38] Instruction about concepts as well as procedures can lead to increased procedural skill.[39] And increasing conceptual knowledge leads to procedure generation.[40][41] However, this relationship is not unidirectional. Conceptual and procedural knowledge develop iteratively, but the conceptual knowledge may have a greater influence on procedural knowledge than the reverse.[41][42] Conceptual instruction led to increased conceptual understanding and to generation and transfer of a correct procedure. Procedural instruction led to increased conceptual understanding and to adoption, but only limited transfer, of the instructed procedure.
Technical uses of the phrase[edit]
Artificial intelligence[edit]
In artificial intelligence, procedural knowledge is a type of knowledge that can be possessed by an intelligent agent. Such knowledge is often represented as a partial or complete finite-state machine or computer program. A well-known example is the procedural reasoning system, which might, in the case of a mobile robot that navigates in a building, contain procedures such as "navigate to a room" or "plan a path". In contrast, an AI system based on declarative knowledge might just contain a map of the building, together with information about the basic actions that can be done by the robot (like moving forward, turning, and stopping), and leave it to a domain-independent planning algorithm to discover how to use those actions to achieve the agent's goals.