PSYCH 644 Week 4 Knowledge Representation and Language Example

How Experts and Novices Store What They Know: Knowledge Representation, Language and Rewriting Work Instructions for a Bilingual Machine Shop

[Student Name]

University of Phoenix

PSYCH/644: Psychology of Learning and Cognition

Week 4 Assignment

[Instructor Name]

[Date]

The aerospace shop, its machinists and the work instructions are composites written for a model paper; research findings and the writing standard come from the sources listed.

Knowledge is not stored as a list of facts but as organized structures that shape what we notice and how we interpret language. When a writer and a reader have different structures, the same words can produce different understandings. This paper applies research on knowledge representation and language to a costly misunderstanding at my workplace.

The Misunderstanding

At Saguaro Precision, each job comes with written work instructions. One instruction for a titanium bracket read: "Drill 0.250 thru, cbore 0.438 x 0.25 dp per dwg, deburr." Experienced machinists read this as drilling a through hole, then counterboring to a diameter of 0.438 inch and a depth of 0.25 inch as shown on the drawing. An apprentice in his third month read "0.25 dp" as applying to the drilled hole and the counterbore together and set the counterbore depth from the wrong reference surface. Forty parts were ruined before inspection caught the error. The apprentice is a fluent English speaker who grew up speaking Spanish at home in Nogales, as do about half our machinists.

How Experts and Novices Organize Knowledge

Chi et al. (1981) asked physics experts and novices to sort physics problems into groups. Novices grouped problems by surface features, such as whether they involved inclined planes or springs. Experts grouped them by the underlying principles needed to solve them, such as conservation of energy. Experts' knowledge was organized around principles connected to methods of solution, while novices' knowledge was organized around literal features of problems.

Experienced machinists read an instruction through a rich structure: they know what a counterbore is for, how depths are referenced on drawings and what makes sense for a bolt head to sit flush. The abbreviations activate that structure, and ambiguous wording resolves automatically. The apprentice, with a thinner structure, relied on the literal sequence of words.

Concepts in a Network

Collins and Loftus (1975) pictured meaning as a web: each idea is a knot, and threads of different thickness tie it to related ideas. Touch one knot and the tug travels outward along the threads, weaker the farther it goes and fading after a moment. The structure of each person's network depends on experience.

For an experienced machinist, "cbore" is strongly linked to "bolt head," "flush" and "depth from top surface," so reading it activates the right reference. For a novice, the links are weaker or missing, and other associations, such as "depth" linked to the drill operation just mentioned, may dominate.

Building Meaning From Text

Kintsch (1988) proposed the construction-integration model of comprehension, in which readers first construct a loose network of propositions from the text and from associated knowledge, including irrelevant associations, and then integrate them through a process that strengthens coherent elements and suppresses incoherent ones. The result is a representation of the text and a situation model of what it describes. Comprehension depends heavily on prior knowledge, which supplies the connections needed for coherent integration.

Our instruction relies on readers to supply nearly all the meaning. Its abbreviations and compressed syntax leave many possible constructions, and only readers with the right knowledge integrate them correctly.

The instruction was written in the language of people who no longer need it.

Language and Bilingual Readers

Kroll and Stewart (1994) proposed the revised hierarchical model of bilingual memory, in which words in each language connect to a shared conceptual store with different strengths. For people who learned a second language later, second-language words are initially linked to concepts mainly through first-language translations, and direct links strengthen with proficiency. The model has been refined and debated, but it captures a common pattern: processing a less dominant language may take longer and rely more on translation.

Our bilingual machinists vary. Many are fully fluent in English and learned shop vocabulary only in English. Others think through technical problems partly in Spanish. Technical abbreviations add a third layer: they are neither ordinary English nor Spanish but a specialized code. Simply translating instructions into Spanish would not help those who learned the trade in English.

A Controlled Language Approach

Aerospace already has a tool for this problem. The ASD-STE100 Simplified Technical English specification, developed for aircraft maintenance documentation, restricts vocabulary to approved words with single meanings, limits sentence length and requires one instruction per sentence and active voice. It was designed to help readers whose first language is not English.

Applying its principles, the revised instruction reads: "1. Drill a hole 0.250 inch in diameter through the part. 2. Counterbore the hole to 0.438 inch diameter. Measure the counterbore depth from the top surface. Make the depth 0.25 inch. 3. Remove all burrs." A small diagram shows the counterbore depth reference.

Schemas and Scripts on the Shop Floor

Beyond individual concepts, experienced machinists carry schemas, organized knowledge about typical situations, and scripts, knowledge of typical sequences of events. A machinist's script for a counterbore includes checking the drawing's depth reference, setting the tool length, cutting a test piece and measuring before running the batch. When a script is well established, it fills gaps in written instructions automatically. Apprentices have partial scripts, and a missing step, such as measuring the first piece, can lead to an entire batch being ruined. Building scripts is part of training, but instructions that state the sequence explicitly protect learners while their scripts develop.

Why Experts Write Compressed Instructions

The original instruction was written by a senior manufacturing engineer who has machined parts for thirty years. Experts often underestimate how much their writing depends on knowledge readers may not share, a pattern sometimes called the curse of knowledge: once something is known, it is hard to imagine not knowing it. The engineer was not careless; his compressed style is efficient for readers like him. The solution is a process that brings novice readers into the review of instructions, rather than relying on writers to imagine what novices need.

Testing the Revision

Rewriting is not enough; readers must test it. Eight machinists took part in a quick check: four veterans and four apprentices, half of each group more at home in Spanish. Each read both versions and told me, in their own words, exactly what they would do at the machine. All eight interpreted the revised version correctly; three of four apprentices misinterpreted the original in some way. Experienced machinists said the revised version took slightly longer to read but was clearer.

Recommendations

The recommendations target both the texts and the readers.

Rewrite high-risk instructions using controlled language principles, add diagrams for depths and datum references, keep a shop glossary in English and Spanish linking abbreviations to plain terms and images and pair apprentices with experienced machinists to talk through new instructions before running a job.

Conclusion

Experts and novices store knowledge differently, and that difference shapes how they understand compressed technical language. Research on knowledge organization, spreading activation, text comprehension and bilingual processing explains a costly misreading and supports clearer, tested instructions that do not depend on readers already knowing what the writer meant.

References

Chi, M. T. H., Feltovich, P. J., & Glaser, R. (1981). Categorization and representation of physics problems by experts and novices. Cognitive Science, 5(2), 121-152. https://doi.org/10.1207/s15516709cog0502_2

Collins, A. M., & Loftus, E. F. (1975). A spreading-activation theory of semantic processing. Psychological Review, 82(6), 407-428. https://doi.org/10.1037/0033-295X.82.6.407

Kintsch, W. (1988). The role of knowledge in discourse comprehension: A construction-integration model. Psychological Review, 95(2), 163-182. https://doi.org/10.1037/0033-295X.95.2.163

Kroll, J. F., & Stewart, E. (1994). Category interference in translation and picture naming: Evidence for asymmetric connections between bilingual memory representations. Journal of Memory and Language, 33(2), 149-174. https://doi.org/10.1006/jmla.1994.1008