Re: Grok or: Information, Knowledge, and Mental Models

Jack Shedd has an insightful piece on his blog concerning shared patterns of thought and language:

If you ever meet two folks who collaborate well, who can finish each other’s thoughts, chances are they share a pattern language. ... Recognizing that our language is not absolute, that labels are open to personal interpretation; Slap whatever label you’d like on it, I’ve found no better way to think of it than in the term of patterns.

This is an issue thats widely discussed in Psychology and Philosophy. It comes down to the difficult relationship between language and thought, and between knowledge and information. Consider Jack's figurative example:

The best way for me has always been to repeat back whatever idea I hear in my own words and try hard only to use terms I know I share with the person. If someone says they want a “patriotic” logo, I immediately say, “So something with red, white, and blue, maybe stars or stripes in it?” Maybe that’s what the person was thinking. Maybe not. He might have been thinking of something airy and old, with black-letter type and dark brown hues. That could be patriotic to someone who thinks of the Constitution, and not the flag, as patriotism itself. He’s no more wrong or right than I am in my definition.
But oh, what a dick I’ll look like when I turn in that blue logo with the star. The client will think he’s chosen the wrong designer, that I didn’t understand his business at all; Worse, that I didn’t listen to him. That’s the first thing everyone thinks when there’s a mismatch. They blame you and think you didn’t listen. Even if you listened perfectly. Even if you took detailed notes. Everyone always thinks it was a lack of effort on your part. That you’re somehow dense, or dumb.

From my perspective, the problem lies in different interpretative frameworks and different cognitive biographies. Both persons (subjectively) know what a patriotic logo is, but their knowledge is highly subjective and contextual. Each person has his or her own "interpretative framework" (Polanyi, 1958). In communication, this subjective knowledge structure is transformed into information that does usually not convey its full set of attributes. The German philosopher Ulf von Rauchhaupt nails it quite well in his definition of knowledge:

The interpretation of information that leads to knowledge by an individual can be seen as an organization process: The process of interpreting data as information is already an act of organization: we perceive the data, order it and link it with other informationfrom our previous knowledge.... In this way, the new information becomes a part of our knowledge for further acts of interpretation. What we know then has a higher degree of organization than the barely obtained information.... Contrariwise, knowledge becomes information again if it is expressed. In order to express knowledge, an individual cannot supply his or her entire network of previous knowledge, which consists of his or her entire history of experiences, his or her cognitive biography. In order to share and exchange knowledge, humans have to partially reduce it to information. The possibility of such a reduction, the possibility to encode knowledge into information, is the reason for the sometimes synonymous use of the terms knowledge and information. Information is a condensed form of knowledge, knowledge is information whose organization exists only for the knowing person (Rauchhaupt, 2005, p. 98f, own translation).

Note that in von Rauchhaupt’s view, knowledge is represented as a network of organized information. Jack has come up with a practical solution: He will verbalize a part of his network of previous knowledge that is connected to the information that he perceives and integrates into his network of previous knowledge. This approach has its limitations, especially if implicit knowledge is involved, but I see it as a good mental exercise that can potentially limit misunderstandings and conflict. Furthermore, I find it quite fascinating that someone arrives at a similar description of the problem based on general wisdom. Furthermore, I think that Jack's illustration nails the issue quite well.

However, I would contradict Jack's original claim that two folks who collaborate well, who can finish each other’s thoughts, share a pattern language. I would say that these two individuals share a task-relevant mental model. In research, such mental models have been operationalized (measured) as knowledge networks (graphs), in line with von Rauchhaupt's concept of a network of organized information. For example, Mathieu et al. (2000) elicited team members' mental model with a structural knowledge elicitation technique similar to my AST.

It turned out that teams in which the resulting knowledge graphs were similar, i.e. in which mental models were similar, performance was higher.

References:

Mathieu, J. E., Heffner, T. S., Goodwin, G. F., Salas, E., & Cannon-Bowers, J. A. (2000). The influence of shared mental models on team process and performance. Journal of Applied Psychology, 85(2), 273–283.

Polanyi, M. (1958). Personal knowledge. London, UK: Routledge & Kegan Paul.

Rauchhaupt, U. von. (2005). Wittgensteins Klarinette - Gegenwart und Zukunft des Wissens [Wittgenstein’s clarinet - present and future of knowledge]. Berlin, Germany: Berliner Taschenbuch Verlag.

UPDATE: Jack sent me an E-Mail and asks:

What is the distinction between a "knowledge network"
and the "pattern language"?

I should have made that clearer. The term "pattern language", as far as I understand Christopher Alexander's use of the term, is a codified (i.e. articulable, speakable) repository of methods for solving typical problems (that occur in design-related areas).

Knowledge can be defined as a set of structural connectivity patterns whose content has been viable for the attainment of goals [1].

Therefore, knowledge is always a network of components that have been referred to differently in different fields and by different scholars, but the network organization of knowledge is something that most people can agree upon.

There are similarities between the concepts pattern language and knowledge, but the difference I see between the concepts lies in the dependency of a pattern language on a codifiable language (and possibly in its dependency on a specific field of expertise). This dependency on codification becomes quite clear in Wikipedia's notion of a pattern in this context [2]:

"A single problem, documented with its best solution, is a single design pattern. Each pattern has a name, a descriptive entry, and some cross-references, much like a dictionary entry. A documented pattern must also explain why that solution is considered the best one for that problem, in the given situation."

In a pattern language, a pattern is thus something explicit and almost objectively true. Knowledge on the other other hand relies in no small parts on implicit content and connections, i.e. things we know but cannot articulate, _especially_ in the context of doing something. Consider these two sentences (example by Wittgenstein):

1. I know the height of the Mont Blanc.
2. I know how a clarinet sounds.

Both sentences constitute knowledge and both pieces of information have an entire network of associated information to them in the head of the person who says either of these sentences. However, neither the associations with these pieces of information, nor the associated items need to be explicitly available, i.e. speakable. We know more than we can say [3]. The network associated with the second sentence cannot be made explicit in such a way that a receiver who has never heard a clarinet before acquires the same knowledge as the sender.

This issue taps in the fundamental question on the relation between language and thought. Without going into too much detail, my point is: Similar knowledge structures between two individuals can lead to the point that those two individuals share a pattern language, but that must not necessarily be so, because parts of the knowledge structure cannot be articulated. However, similar knowledge structures condition successful cooperation in task performance (at least thats what some scholars claim; others, like myself, argue that a certain amount of cognitive heterogeneity between team members facilitates higher team performance).

Thus, the configuration of knowledge structures conditions the ability to work together, but the link between knowledge structures and pattern languages is not 100%, i.e. there can be diverging languages that base on similar knowledge (structure/networks). It is possible that two people can have similar knowledge, can work together, but have a different pattern language.

References

[1] Meyer, B., & Sugiyama, K. (2007). The concept of knowledge in KM: a dimensional model. Journal of Knowledge Management, 11(1), 17–35.

[2] http://en.wikipedia.org/wiki/Pattern_language#What_is_a_pattern.3F

[3] Polanyi, M. (1983). The tacit dimension (Reprinted ed.). Gloucester, MA: Smith.