Systems Analysis

This is individual work, ongoing through out the semester. Roughly 250 words per note, submitted on the course site before the meeting the reading supports. Credit / no credit. There are nine readings; seven of your reading notes must earn credit.

The prompt for each note is the same all semester:

Name one thing in this reading you think is wrong, or one place where our situation doesn’t fit what the author describes. This is not a summary. If you agree with everything in the reading, you haven’t read closely enough.

Notes that summarize get no credit. Two notes can be missed entirely, or earn no credit, with no penalty. Below seven, each shortfall costs one seventh of this component of your grade.

Group deliverable, submitted on this site. Not graded, but important: both finding-out portfolios, the relevant systems your group commits to on 10/7, and the models built from them are all affected by this statement.

In four or five sentences:

Name the aspect of the merger your group is inquiring into for the rest of the term:

1. curriculum
2. advising and student services
3. student organizations
4. internships and field experiences
5. external communications
6. internal communications

… or one of your own design, as long as students can see into it, and if the people who know about it would talk frankly to an undergraduate. Faculty hiring, tenure cases and budget lines fail the second test.

Say what your survey of potential sources found on it.

Say what the survey failed to find that you expected to be there.

Name the roles of the people who would have to talk with you for this inquiry to go further than just documents.

Your group’s first rich picture of our situation, drawn by hand in the room and photographed.

Graded on completion only. Rich pictures are instruments of thought, not products. Grading their quality would teach you to draw pictures that look like good rich pictures, which is the opposite of the point.

Photograph the picture at the end of every working session and date the file. The iteration memos require you to reproduce an earlier artifact and say what was wrong with it, which is impossible if the only rich picture that exists is the current one.

Group deliverable, in two parts.

Analysis One — your inquiry. Who are the clients, who are the practitioners, and who are the issue owners in your group’s inquiry into the aspect you committed to? The list of issue owners is a choice, not a discovery, and different choices open different inquiries. Say what your choice opens and what it closes off, including who is affected but not consulted.

Then one paragraph on the choice of aspect: what your aspect made available to this analysis, and what taking a different aspect would have put on the issue-owner list that is not on it now.

Source inventory. The documents your group has gathered, each with its date and its audience noted. Public announcements, org charts, strategic plans, curricular filings, faculty governance minutes, news and student-newspaper coverage, alumni communications, etc.

Sources you gathered before committing to your aspect belong here too, whether or not they bear on it: what you gathered and could not use is evidence about where this school writes things down.

Assessed on: whether the three roles (client, practitioner, issue owner) are argued as choices rather than reported as findings; whether the affected-but-not-consulted question is answered with parties specific enough to say what they would want; whether the inventory records date and audience rather than just listing links.

Group deliverable, in two parts: an Analysis Two of about 1,000–1,200 words, and an Analysis Three of about 500–800. Analysis Three is shorter because it is the smaller analysis, not because it matters less. Write each part from its own dated file (the Analysis Two file you have kept since 9/14, the Analysis Three file you keep from 9/28) rather than reconstructing it the week it is due. Where a claim rests on a particular entry, give the entry’s date.

Analysis Two — the social situation

What to write. The roles present in your aspect of the situation, formal and informal; the norms of behavior expected of whoever holds each role; and the values by which performance in a role is judged good or bad. The informal roles often tell you the most: the person everyone actually asks when the official process fails.

Show how the three elements (roles, norms, and values) affect each other. Checkland’s point is that none of them stands alone. Holding a role changes the person who holds it; how they hold it shifts what is expected of the next holder (norms); and that shifts what gets praised (values).

Where to look. Read the norms from practice, not only from what official statements say they are. The most useful material is in two places: the gap between stated and enacted norms, and the places where the two divisions, DILS and DDSS, expect different things of the same nominal role. Practice is not only what people tell you. A document recording what was done, compared to one saying what would be done, shows it too.

Analysis Three — the political situation

What to write. The commodities that signal power in your aspect, and the processes by which they are obtained, used, protected, defended, passed on and relinquished. A list of commodities is half the analysis; the processes are the other half. Start from the page posted before 9/28: say which of its candidates are relevant to your aspect, which are not, and what is missing from it.

What to explain. A good Analysis Three explains why an actor in a given position would act as that position’s incentives suggest, whoever occupies it.

One rule for both parts: roles and positions, never persons. Naming people turns a structural analysis into gossip. A role or position with just one holder is an identifiable person; describe it by the anonymization rule the class set on 9/14.

Assessed on: Analysis Two — whether the norms are read from practice; whether roles, norms and values are shown affecting each other rather than simply listed. Analysis Three — whether it sorts the commodity page into relevant, not relevant and missing; whether it names processes and not only holdings; whether it explains why whoever holds a position would act as that position’s incentives suggest. Both — the no-named-individuals rule.

The second version of your group’s rich picture. v1 was drawn in the room before your group had an aspect, and pictured the merged school. v2 pictures your group’s aspect: the social and political relations your Analyses Two and Three have found there.

Graded on completion only. Keep the dated photograph of v1—you will need it.

What to draw. v2 is due the same day as Analyses Two and Three, and that is deliberate. A written analysis sets out roles, norms, values and commodities one at a time; a picture is where you can see how they bear on each other—which role holds which commodity, which norm protects which process, where two sets of values meet and do not agree. Draw those relations. Positions and processes, never persons: the rule that governs Analysis Three governs the picture of it.

How. There is no class time for v2. Draw it together outside class, by hand or in a shared digital canvas such as Excalidraw. What matters is that the group draws it, not that one member draws it for the others. Start a new sheet rather than drawing over v1, and photograph it at the end of every working session.

Three times this term, in ≤ 800 words:

Identify one specific artifact you or your group produced earlier — a rich picture, an Analysis One list, a root definition, a comparison row. Quote or reproduce the relevant part.

State what you now think is wrong or incomplete about it.

Identify what specifically caused the change in your view: a document, a conversation, a classmate’s model, a moment in a debate. Be concrete. “Reflection” is not a cause.

State what you changed it to, and what was lost due to that change — what the earlier version made visible or possible that the new one does not.

Assessed on: traceability to specific artifacts and specific causes; whether the change is substantive rather than cosmetic; whether the cost is acknowledged.

These four paragraphs are required, not suggestions. “I gained a deeper understanding of the situation” is worth nothing, because it identifies no artifact, no change, and no cause.

Group posting to Zulip, in #general, topic “pqr statements”, not to this site. Not graded.

Write down the two or three relevant systems your group committed to in class on Wednesday 10/7, all within the aspect you committed to on 9/14. For each system, give four things:

1. The PQR statement, as one sentence: do P, by Q, in order to contribute to achieving R.
2. The worldview that makes the system worth having, in a sentence or two: who would think it worth doing, and what they would say it is for. Positions, never persons.
3. Primary-task or issue-based, with the reason in a phrase: which organizational boundary the system takes as given, or which it cuts across.
4. Whether anyone in your group holds this worldview. Yes or no.

Across the two or three, make sure that:

• The worldviews contrast. Two systems that the same people would praise for the same reasons are really just one system written twice.
• At least one is a worldview nobody in your group holds.
• Both kinds are there: at least one primary-task system and at least one issue-based.

Post to Zulip, in #general, topic “pqr statements”, so that every group can read every other group’s systems. One message per group, posted by the recorder and headed with your group number and aspect. If you revise a statement, post the revision as a new reply; do not edit the original.

A worldview nobody in your group holds will be read by classmates who might hold it. Write it in such a way that they would agree it represents their view.

R implies a worldview, and the transformation is represented by Q (the declared how), not by P. A group that can state its P and its R but handwaves at Q does not yet have an identifiable transformation. Q also has to be a plausible how for P: if P is “define health-care needs” and Q is only “by asking patients for their views,” Q cannot be defended (Ch. 2, p. 39).

A relevant system is a choice of viewpoint, not a part of the organization, and one model expresses exactly one worldview. A model that tries to be balanced argues for nothing and can be debated with by no one.

Your aspect is the subject matter: what your group is inquiring into. A relevant system is a viewpoint on that subject matter, and you are generating several. Do not answer “what is your relevant system?” with the name of your aspect.

Not assessed. Like the inquiry focus statement, this is a decision that future graded work depends on. On Monday 10/12 you will turn each of these systems into a root definition and CATWOE. Before Monday I will check: whether every statement has a P, a Q and an R, with Q specific enough to represent the transformation and plausible as a way of doing P; whether the worldviews contrast; whether at least one is held by nobody in the group; whether both kinds of system are present; whether any system is just the name of the aspect; and the no-named-individuals rule. A group whose statements fail a check hears so in the thread before Monday, and should fix them there.

Individual, posted to Zulip in #general, topic “steelman declarations”, by 6 p.m. Tue 10/13. Not graded.

Declare which worldview you will take on for the steelman exercise. It must be one you do not hold and are not sympathetic to.

Post to Zulip, in #general, topic “steelman declarations”, one message, by 6 p.m. on Tuesday 10/13. State the worldview the way the PQR statements did: who would think something worth doing, and what they would say it is for. Positions, never persons. The PQR thread and the three disagreements on the board from 10/7 are a place to start, but any worldview at work in our situation will do.

Then read the others. If you hold a worldview someone else has declared, reply to their message to say so, by Tuesday evening. Those replies are how readers are found: your steelman will be read by classmates who hold the view, and I assign readers from this thread on Tuesday night. If nobody replies to yours, I will ask you that night to take another view, such as the idea of fairness most of us argued from on 8/26.

This is not administrative. The steelman is validated by classmates who do hold the view signing off on your account, so a worldview nobody in the room holds cannot be validated, and a worldview everybody picks leaves no one to validate it. Declaring in advance lets the distribution be rebalanced before you write, and declaring in public is what finds the holders.

Individual, ≤ 1200 words. Handed in on this site by 11:59 p.m. on Wed 10/21, and sent to your readers on Zulip the same day.

Take the worldview you declared on 10/13, or the one I asked you to take instead: one at work in our situation that you do not hold and are not sympathetic to.

Write a root definition and CATWOE expressing it. The worldview must be stated as its holders would state it — as something a reasonable person could believe for good reasons — not as something to be explained away by interest or ignorance.

Then state, in one paragraph, what this worldview sees that yours does not.

Your account will be read by classmates who hold this view. The question they will be asked is: could you sign this?

Hand it in on the site by 11:59 p.m. on Wed 10/21. The same day, send it to your readers, listed in #general › steelman declarations: one Zulip direct message to them and to me together, with the file attached.

If you are a reader, answer in that message by 11:59 p.m. on Fri 10/23: yes, you could sign it, or no, quoting the sentence that stopped you.

Assessed on: peer recognition as the primary criterion; absence of tells; whether the final paragraph identifies a genuine blind spot in your own position.

The tells, named in advance. Scare quotes, “they believe that…”, explaining a position by the interests of those who hold it, and any sentence that would make a holder of the view wince. Each of these costs points under absence of tells.

Individual. Built alone during the asynchronous week — this is the competence check, and the only instrument in the course that asks one student to produce a whole systems-thinking-stream artifact unaided.

Start from one of your group’s system cards, one whose worldview you hold; the one your group modeled in class on Mon 10/19 will do. Change one element of its CATWOE, usually W, though O or E will also do. Say which card you started from, what you changed, and what else changed with it. The root definition that results is the one you model, and it should be yours alone: if someone in your group has made the same change, one of you changes something else.

Four parts:

• a root definition;
• CATWOE, with the transformation as the hard part: T operates on one entity, converting it from one state to the same entity changed. A T with different input and output entities is the most common error and does not pass;
• an activity model — seven activities give or take two, ordered by logical dependency rather than chronology, with a monitoring-and-control subsystem;
• criteria for efficacy, efficiency and effectiveness, stated separately, because they are independent of one another.

Model from the root definition and nothing else. Importing what actually happens in the school collapses the two streams and produces a comparison of X with X, which can teach nobody anything.

Hand in one PDF, in this order: the card you started from and what you changed; the root definition; CATWOE; the activity model, with every activity numbered and the root definition at the top, as a photograph of a paper sheet or an Excalidraw export; and the three criteria. Hand it in on this site by 11:59 p.m. on Fri 10/30. The same day, send it to your two critics, listed in #general › model set critics: one Zulip direct message to them and to me together, with the PDF attached.

A model is defensible, not correct. Two competent modelers working from the same root definition produce different models, because they read the same words slightly differently. The test is whether you can defend yours as representing your root definition and CATWOE.

Assessed on: whether T operates on one entity; whether every phrase of the root definition leads to something in the model and every activity links back; whether the monitoring subsystem is present and its three criteria genuinely distinct; whether the model is defensible from its own root definition rather than correct about the school.

Individual. Required, ungraded.

On Fri 10/30 two classmates will send you their model sets, each in a Zulip direct message to you, another critic and me. Your two are listed in #general › model set critics. Answer each one in the message it came in, by 11:59 p.m. on Sun 11/1, against the two questions from class on Mon 10/19:

1. Does every phrase in the root definition lead to something in the model?
2. Can every activity in the model be linked back to the root definition or CATWOE?

For each thing you find, quote the words in the root definition that lead to nothing, or name the activity, by its number, that links back to nothing. If you find nothing for a question, say so. “I’d add…” is a preference, not a finding. If a set has not arrived by Saturday, send its author a Zulip direct message, with me on it, saying so. I don’t answer on weekends and will take it up on Monday; a critique of a set that never arrived is not late.

Three times this term, in ≤ 800 words:

Identify one specific artifact you or your group produced earlier — a rich picture, an Analysis One list, a root definition, a comparison row. Quote or reproduce the relevant part.

State what you now think is wrong or incomplete about it.

Identify what specifically caused the change in your view: a document, a conversation, a classmate’s model, a moment in a debate. Be concrete. “Reflection” is not a cause.

State what you changed it to, and what was lost due to that change — what the earlier version made visible or possible that the new one does not.

Assessed on: traceability to specific artifacts and specific causes; whether the change is substantive rather than cosmetic; whether the cost is acknowledged.

These four paragraphs are required, not suggestions. “I gained a deeper understanding of the situation” is worth nothing, because it identifies no artifact, no change, and no cause.

The final version of your group’s rich picture, redrawn in light of everything since v2.

Graded on completion only.

How to hand it in. Upload the picture itself: a photograph, an exported image or PDF, or a screenshot of the whole canvas. A link to Canva, Excalidraw or another service is not a submission, because it shows whatever the canvas holds when someone opens it, not the version you handed in.

The situation is not what it was in August, partly because we have been inquiring into it. That is not a methodological failure; it is the condition of all such inquiry.