Sei sulla pagina 1di 8

HIERARCHAL NETWORK MODEL 1

Running Head: HIERARCHAL NETWORK MODEL OF SEMANTIC MEMORY

Hierarchal Network Model Of Semantic Memory Theory Critique

MaryBeth Davis

00016225

College of Science Technology and Applied Arts of Trinidad and Tobago

Lecturer: Roslyn Humprey

PSYC 468: Cognitive Psychology

HIERARCHAL NETWORK MODEL 1

Semantic memory is defined as general knowledge that does not require memories of specific events in which one was supposed to have learned this information, it is really the what of memory. The beginning of research on semantic memory was on artificial intelligence and machine translation which began in the late 1960s. Researchers Collin and Quillian proposed a model called hierarchical network model as they were intrigued with efficiency of storage in memory in humans and to the fact that the model can be shared between A.I and humans. The hierarchical network model is a theory of semantic memory by Collin and Quillian of 1969 that is to be critiqued according to the guidelines of what makes a good theory. (University of Minnesota , 2010)

The hierarchal organization was utilized in studies because of its ability to facilitate the recall of semantic information which was confirmed by research done by Bower, et al (1969) on the recall of hierarchal information. Which concluded that subjects do organize input into hierarchies as it allows one to structure memory in such as way to make searching more efficiently. (San Diego.edu)

The hierarchical model assumes that category information is stored directly in memory by means of associations. The properties /features true to all properties /features are stored at the highest level and basic level categories are stored on the intermediate level and the lowest level stores are properties/ features that are true for that particular member but not all in that category. The hierarchical theory has cognitive economy as information is never repeated at each level in this memory store system, but this may make retrieval more complicated because access to more than one level may be needed to retrieve the necessary properties /features to decide category membership. The main tenets of the theory is that it takes time to move from one level of the

HIERARCHAL NETWORK MODEL 1

hierarchy to another, and additional time is required to retrieve the features stored at any one of the levels. These tenets were verified through a demonstration with sentence verification were they wished to show that it takes longer to verify a sentence that has two components from a different level and it takes longer to verify sentences when features need to be retrieved. The hierarchical model is shown by two sentence verifications that have information about a category at the same level; verification of the second is faster because activation caused by verifying the first sentence corresponds to initial activation of the level. Sentences that require information from different levels dont have this consideration. (San Diego.edu)

Activation of memory causes information that is correlated to the node in the network and the other connected related nodes to be primed. The activation travels through the links that connect the network of nodes which are interconnected. Therefore the memory retrieval of semantic data is first gotten by entering the network at the concept/node that corresponds to the item trying to be retrieved, then examination of the node by way of its properties, a match is found we stop if not continue one step in the hierarchy and continue to examine that nodes properties until a match is found. The hierarchical model posits that relation between two categories is stored in memory in a semantic network. The semantic network is where concepts/ nodes are joined by other concepts by links. The links specify the relation between concepts/nodes and also that activation of a related concept/node spreads that activation along the links in the networks to its related concept/s/ nodes. This explanation or theory is used by cognitive psychologist as a way to explain the organizational and retrieval of information in long-term memory. (University of Minnesota , 2010)

HIERARCHAL NETWORK MODEL 1

human ^Figure 1- example of hierarchal model with nodes & links | | politician --is--> in party y | ^ is-a | | is-a | | | | Ms. X ----is--------> MP ---can---> vote in p.

HIERARCHAL NETWORK MODEL 1

This theory has received criticism as a great number of research findings have debunked the hierarchical network. There is the problem that not all categories can be structured in a hierarchy as abstract concepts bring challenges. The information processing task research results is one such study but one must keep in mind that the hierarchal model is a connectionist approach and that the connectionist approach was an answer to information processings serial processing of the data. Another flaw was found in the sentence verification study were the material was found to be faulty , the size of the sample category and the words used in the study by Collin and Quillian confused the semantic distance with conjoint frequency( when two words with a clear association occurs together frequently). This has caused critics to conclude that the results of the study did not support the hierarchal model as it was due to typically of words and semantic distance distortion.( Harley,1995) These criticisms caused Collin and another colleague Loftus to address these factors by revising the hierarchical model where the role of hierarchal relationships is minimised. (Reference.com, 2008) Comprehensiveness is how well the theory explains a large number of observations and this theory in the beginning of its inception debunked the behaviourist notion of meaning being derived from a network of associations as they viewed the placement of a word in the network of associations as the words definition. Episodic not semantic instances gave meaning and this associative nature of meaning does not entail all the features of a word. The hierarchical model answered the flaws in the behaviourist theory as it provided structure, relationships between words and cognitive economy. Therefore at inception it explained observations of memory, organization of stimuli from the environment and learning that other theories lacked but later on it itself was debunked due to flaws of its own.

HIERARCHAL NETWORK MODEL 1

Internal consistency is based on whether the assumptions and propositions of the theory fit together in a coherent way. In the beginning of reading about the theory you can note that the theory is very logical in its explanation of how memory is organized in semantic memory. The explanation of there being concepts that are called nodes which form a network that includes other nodes all joined together by links which activate nodes that has the same categorization and properties of the word that activated the retrieval process and the main fact that retrieval is done by levels as from highest to lowest in a hierarchical nature are all logical. It is the critics of the theory who brought to the forefront the issue that the basis of the theory is based on faulty methodology by the experimenters. A theory is really based on assumptions and propositions as this defines it and since the hierarchical models significance is questioned and showed to be lacking, the assumptions and propositions that defined it based on the sentence verification study lack internal consistency. The theory was revised and the hierarchal nature of semantic memory was minimized by one of the original theorist Collin which shows that it originally lacked internal consistency. Heuristic function is the ability of theories to stimulate further research. In the critique of hierarchal network model complied by Hartley (1995) he showed why criticisms arose due to the fact that other research based on hierarchical nature of semantic memory provided evidence that showed that the proposed theory did not work that way . These researches discredits the validity of the theory but shows that further research was stimulated by it but it only proved to invalidate its assumptions and propositions. The parsimony or simplicity of the theory is shown by the experiment of sentence verification studies done and the ease that other researchers had in duplicating and rebuking the theory. The hierarchal network model has no functional significance due to the fact that the assumptions and

HIERARCHAL NETWORK MODEL 1

propositions were wrong on how learning occurred but if we were to go back to its inception and its rebuking of the behavioural theory of associations one of the reasons for its acceptance one can infer is that it filled in gapes that the behavioural theory had on how learning occurs and therefore was understandable by the average person. As in order for it debunk one theory there must have been a gap that it filled and it can be the ability for persons to fully grasp a subject whose preceding theory did not paint a full picture that everyone could understand. The theory cannot be applied to any setting of learning as it has been debunked and revised in such a way that the core posit of hierarchal structure of semantic memory was given a walk on role in the revisions. One should note though that the hierarchal network was debunked the hierarchal organization of data to facilitate the retrieval of data from semantic memory stilled showed that people do organize information in a hierarchy of categories that makes retrieval easier than no organization. This is shown in learning as difficult information is broken down into a hierarchy in order to simplify it. The Hierarchal network model (1969) is comprehensive, it filled a gap in how memory works/retrieved that was needed at the time and has facilitated research on semantic memory and how it is organized which is what a theory does. It has no internal consistency, functional significance and cannot wholly be applied to learning but even with these faults it is easy to understand as it has parsimony.

HIERARCHAL NETWORK MODEL 1

References

Harley, A. T. (1995). The psychology of Language: From data to theory. East Sussex,UK: Psychology Press. Reference.com. (2008, May 05th ). Collins and Quillian Semantic Network Model. Retrieved November 15th, 2011, from Reference.com: http://www.reference.com/browse/Collins+&+Quillian+Semantic+Network+Model San Diego.edu. (n.d.). Semantic Organization. Retrieved November 15th, 2011, from http://home.sandiego.edu/~taylor/semorgn.html University of Minnesota . (2010, April 03). Semantic Memory. Retrieved November 15, 2011, from Psych 5054: http://www.psych.umn.edu/courses/spring06/fletcherr/psy5054/semantic.html Figure 1 attained from: Edu tech Wiki. (2009, August,01). Semantic Network. Retrieved November 15th,2011,from Edu tech Wiki: http://edutechwiki.unige.ch/en/Semantic_network

Potrebbero piacerti anche