First Part

Topic One: Training

What is the role of training in organizational learning?

Training is the method to pass organizational memory and routines from one person to another. It helps maintain organizational memory in the organization and socialize new members. It also transmits needed skills, values, etc. to the entire organization,or targets skills in certain members.

What are organization-specific skills?

Organization skills are those that cannot be acquired from outside. These include:

* values, beliefs,

* internal procedures, rules

* organizational communication channels

* proprietary technologies

(Similar to context-specific knowledge seen in consultancy article).

If there were no such skills, why would an organization engage in training of personnel?

One reason is to ensure that the appropriate people have the right mix of skills for their role. Another is to socialize members through training exercises. Training can help build a common skill base in the organization. It also may be more efficient to do the training internally (one cannot expect everyone you hire to possess all the needed skills).

What would happen if training were two-way, that is, if the trainers learned from the trainees as well as vice-versa?

I think that trainers do learn from trainees already, through the act of deciding what to teach reinforces and expands their own understanding. Overall training effectiveness for the organization would be increased, and ultimately the organization could learn faster. Also trainers would be more willing to train, and trainees may be less restrained in asking for training. It would be a win-win for everyone.

Topic Two: Hypercycles of Skill

What is the basic problem to which Padgett's model is addressed?

Padgett wonders how skills are reproduced within an organization. What is the process by which knowledge and learning is "reproduced" and thus maintained in a organization where people are constantly coming in and out of it? Why are some skills forgotten and others constantly reproduced? How do different organizational ecologies induce different degrees of complexity in skill mixes of individuals?

What is his approach to the problem?

He borrows a theory from biological chemistry called hypercycles to theorize about how skill mixes are maintained in organizations. He believes that the organizational characteristics produce self-reinforcing cycles of interactions between individuals that maintain certain skills within the organization.

Why does he use the word "reproduction" rather than learning? Does it matter?

He uses reproduction to partly tie it back to its biological roots, and partly as a simplification. Learning is an increase of knowledge or acquisition of a skill -- but his models don't necessarily increae organizational learning (in fact he holds knowledge constant).

Reproduction mainly refers to skills, while learning often refers the increase in knowledge. However, the acquistion of a skill is in fact learning (as well as the loss of a skill) at the individual level. He also uses reproduction to emphasis the transmission of sklls (like genes) in an organization.

His use doesn't necessarily matter at the individual-to-individual level, but there is certainly a distinction at the organizational level.

What is the distinction between "source-only" and "target-only" reproduction?

In "source-only" reproduction, only the source adds the other skill (the target completes the interaction but doesn't learn the other skill). In the "target-only" reproduction only the target acquires a new skill.

What is the result? Does it matter?

He found that in the simplest case (two elements), the target-only reproduction produced stable hypercycles of skill reproduction that were more complex than source-only systems. Usually there were multiple-skilled people at the core and single-skilled people at the edges. Interestingly, stable systems were found above 5 elements in spatial but not non-spatial systems. An organizational structure seems necessary to get complicated skill sequences to emerge and be stable.

In the source-only case everyone was specialized, and the clusters were much smaller (it breaks down to self-reinforcing dyads). Eventually the whole system died out.

Interestingly, only target-only systems maintained the total number of skills. The joint system maintained them only by chance.

Ultimately, this model shows that altruistic learning helps maintain organizational memory because it ensures proper reproduction of needed skills.

Topic 3: The Spread of Knowledge:

What is the basic idea of diffusion?

Diffusion is the spread of something among a population. Things spread from a source with the characteristic and accepted by targets without the characteristic. It assumes people are connected to one another.

What similarities are there among the spread of diseases, the spread of fads, and the spread of knowledge?

All are dependent on communication links, which are often not symmetical or reciprocal.

They require contact, transmission, acceptance, permanence, etc.

What kinds of knowledge will spread fast?

Important knowledge, simple-minded knowledge, high source credibility, high source status. Knowledge with a higher frequency of use will spread faster too. Knowledge that can be adopted incrementally will also spread faster than one that must be taken all at once.

Who will learn what first?

Those for whom the knowledge is consistent with prior beliefs will learn it quicker. Those closer to the initial source or have higher frequency of exposure will learn quicker.

Some also say that some people are more likely to be "early adopters" too.

What is the connection between network analysis in organizations and studies of learning?

Assuming that most knowledge and skills are transferred from person to person, the network of interpersonal communications can help understand how knowledge is diffused, maintained, and increased in organizations. Network analysis also implies that social distance is important, which impacts diffusion of learning as well.

Padgett's model is an example of diffusion of skills through a network.

Topic 4: Fundamentals of the Learning Process:

What would a rational theory of diffusion look like (i.e. if we imagined each organization decides rationally whether to seek or discourage the spread of things)?

The Yale Approach (Hovland) in a sense is a rational approach to diffusion of learning. They focus on the:

* communicator (credibility, similarity, attractiveness, persuasive intent)

* communication (discrepancy, 1-sided, repetition, rational vs emotional

* audience (source) info level commitement initial position, forewarning

* content (delay, reinforcement, group pressure, distraction)

It would also include some network analysis to measure the spread of things in the organization. At the organization level a theory may also include:

* number of initial sources

* desired speed of diffusion

* desired end-point

What would a learning theory of diffusion look like? Other theories?

1. Contact

* frequency of interaction between teacher and learner

* number of "learned" people in network

* interpersonal network

2. Transmission

* teaching effectiveness (infectability)

3. Acceptance

* student's learning ability (reception, susceptability)

4. Diffusion Rate / Diffusion Gradients / Extent of Diffusion

5. Reaction Rate

To what extent should we treat imitation as basic human attribute rather than an attribute to be derived from such assumptions?

In a sense they are both the same. Imitation is also a form of diffused learning.

What are the implications?

According to Cohen & Levinthal, most innovations are borrowed. Getting new ideas from others is often the most effective way to diffuse and maintain knowledge in an organization.

All organized learning is a controlled diffusion process, though in this case there can be "learning reactions" producing new ideas that subsequently diffuse throughout the organization.

Second Part:

What is the dillema in the diffusion of knowledge identified by Cohen & Levinthal?

An organization needs a certain level of absorptive capacity to identify new opportunities and capitalize on them. If they get behind, they won't have the capacity to capitalize on new ideas which will put them further behnd.

Increasing absorptive capacity requires previous absorptive capacity. Competitive R&D requires previous R&D. If you quit investing in R&D, you may never catch up.

What are the assumptions on which it lies?

1. Most innovations result from borrowing.

2. It assumes that a certain level of knowledge is required to identify, acquire, and capitalize on a borrowed idea.

3. Absorptive capacity is path dependent (your current ability is based on previous ability). The more R&D you do the more efficient you are at it.

4. Much of the needed diversity to increase absorptive capacity is in the individuals.

In particular, to what extent is it a dilemma only within an incentive theory of human action?

The other dilemma is that increasing R&D increases the chance of spillover (someone else stealing your idea), but increases absorptive capacity and your ability to capitalize on someone elses spillover. There are thus both positive and negative incentives to increase R&D spending.

Might there be parallels in other visions of action?

This is similar to March's exploration theories that successful firms have slack which allows them more experimentation and thus a better chance for future innovation. In technoligical rich firms, successful firms will become more successful.

There are also parallels to population ecology, where in highly competitive firms the ones most adaptable to the environment (fast R&D, etc.) will survive.

What are the implications for organizational learning?

Firms engage in basic research that has high spillover because it gives them the ability capitalize on state-of-the-art knowledge as well. It shows that organizations cannot sacrifice learning capability for short-term gain without risking their survival.

It also shows how important the learning environment is to organizational action. Absorptive capacity becomes even more important in difficult learning environments.

This reinforces the need to balance exploration with exploitation to stay ahead and prosperous.