On the Relation between Designing and Implementing in Permaculture – Part 22

Okay, we are on the home straight here in what is the twenty-second post in an inquiry into the relation between designing and implementing inside permaculture design process. It is time to somehow sum all these adventures up and bring home the key discoveries. Here goes.

The Standard Permaculture Approach

I started this inquiry by showing that differences aside, every description of a coherent permaculture design process I could find (including my own) treats design and implementation as two separate phases. I mean two separate phases in the sense that you complete the entire design, typically to a high degree of detail, and only then implement it. After Christopher Alexander, we’ve been calling this widespread view of design¹ process a fabricating approach.

Here it is diagrammatically (click here for the key to this and the similarly formatted diagrams below):

Design happens first. Implementation happens second. You come up with the design. You actualise the design.

In a fabricating process the rhythm is decide-draw-decide-draw-decide-draw before moving on to a big chunk of (post-design) DOING

Implementation brings up new information that feeds back into the design, yes. Some writers emphasise this fact less, others more. None deny that such iteration happens. That the design evolves based on what happens as it is sequentially actualised. But this feedback-driven evolution kicks in only after the design has been completed.

Issues with this Standard Approach

In this inquiry I’ve shared some compelling arguments² that this fabricating approach cannot fail to compromise the quality of our design work, and of the gardens or whatever else coming out of that design work.

Indeed, drawing on the work of Christopher Alexander, I’ve tried to show how any attempt to complete a detailed design before implementation involves so much premature and hence arbitrary guesswork and imposition that the quality of the design process is almost necessarily crippled. Though Alexander was mostly interested in buildings, his point applies equally to everything permaculturists design and create. In Alexander’s approach to making things:

…it is theoretically impossible for a successful [thing] to be built from a set of drawings which specify every detail, because that would cripple the capacity of feedback to help shape the elements as they are built (Christopher Alexander 2005, p. 485)

Indeed, as we gleaned from a brief chat with an acorn, the idea of a separate up-front detailed design before implementation flies in the face of how living systems themselves come into being and grow. This is a little bit embarrassing for a design approach supposedly committed to mimicking natural systems!³

These are not the sort of issues to be downplayed, ignored, swept under the rug. They are issues worthy of shining a spotlight on, of bringing out onto the table. Of sorting out.

So where to from here?

Unsatisfactory ways of Resolving these Issues

One approach is to try and patch things up.

We can add more arrows. We can continue propagating idealised linear sequences requiring multiple disclaimers about how non-idealised, how messy, iterative, interrelated the reality of using them is.

I don’t find that approach satisfactory.

Neither is the alternative of what I’ve been calling winging it. While one might use phrases such as “going with the flow”, “being organic” and so on, in this context such phrases are euphemisms for doing shit semi-randomly and generally steering oneself directly into chaos.

The upshot is that we have found both winging it and fabricating to be fundamentally flawed ways of understanding the essence of a sound permaculture design process. Hence the big red crosses. Thumbs down, dude.

Two Promising Leads

The bulk of this inquiry has been an investigation of two alternative framings of permaculture design process. Framings which avoid the issues inherent in both fabricating and winging it. I have been calling these two alternatives the hybrid and the generating approaches, as shown to the right of this diagram:

In exploring this space, in addition to the voices shared below, I owe a debt of gratitude to the action-centric and specifically the agile movements in software development, inside which much development in these directions has been going on for decades.⁴

The Hybrid Approach

In the hybrid approach, which I happened onto thanks to a nudge from Bill Mollison, you complete a concept-level design only before commencing implementation. The detailed design then emerges from within the implementation. As Bill put it in the Designer’s Manual:⁵

Break up the job into small, easily achievable, basic stages and complete these one at a time. Never draw up long lists of tasks, just the next stage. It is only in the design phase that we plan the system as a whole, so that our smaller nucleus plans are always in relation to a larger plan.

In the below diagram the hybrid approach entails a little run of decide-draw-decide-draw up front and then jumps across to decide-draw-do-decide-draw-do for the bulk of the journey (The idea being that you do just enough fabricating to make sure you’re not about to get yourself in trouble, and then it’s generating time):

I shared an example of a hybrid approach starting here.

In this example, we only got as far as this concept design before commencing implementation:

Guess what? It worked a treat. Nothing went wrong. In fact a lot more went right than my hundreds of past experiences trying to make a fabricating approach work. The outcome fits its context beautifully, and really feels like it belongs:

It was also just really nice, to the point of being relaxing, to not have to worry about making detailed decisions ahead of time with a pencil. By making them inside the moment of implementing them, we instead used the bobcat, shovel and rake. This way, each detailed decision was so fully informed by the actual 100% real reality of where the system was at, that it was unquestionably better than we could ever have predicted up front.

Needless to say, I left the experience all but convinced that the way designing and implementing are related is absolutely crucial to the quality of the outcome.

I also left with the question of what would it feel like to start implementing before even a whole site or area concept plan was done.

The Generating Approach

In a generating process, such as that shared here, not even a whole-site concept plan is drawn up before implementation begins.

Here, even more differently to the standard permaculture mantra of:

• observe (people and place or whatever)
• concept design
• detailed design
• implement
• evaluate/tweak

The process, as exemplified in this practical example, is instead something like:

• Immerse in the overall context of the design
• Decide on what high-level features or aspects to tackle first
• Rapidly generate then iteratively test or prototype a first step until something feels solid and relatively certain
• Adaptively implement that step
• Re-immerse in the new reality of the just-transformed whole

Furthermore, all these things end up overlapping in time, with the idea of a linear sequence losing all relevance.⁶

This is the process as used to generate the emerging Mayberry Woodend landscape:

Here the rhythm is decide-draw-do-decide-draw-do-decide-draw-do such that the designing and drawing only get ahead of doing by a decision or two:

In a generating process, apart from the first decision, all key decisions are directly prompted by the just-updated reality of the actual real situation.

Christopher Alexander equates a generating process with an unfolding process, arguing that:

The more one understands the idea of unfolding, and the more one understands the key role which sequence plays in the unfolding process, the more it becomes clear that the process of design and the process of construction are inseparable” (2002, p. 322)⁷

Generative Whisperings from Within Permaculture

Because a generating approach appears at first glance to be the most radical departure from what the permaculture books say, I want to share here some fascinating statements from well-known and respected permaculture design authors in which a generating approach is clearly (to me at least) being hinted at, if not explicitly spelled out as such:

David Holmgren

Master planning, (where detailed plans are implemented producing a final fixed state which is a copy of what is on paper) has been discredited in the planning profession due to its failure to deal with complex evolving systems such as cities. Many attempts at farm planning by consultants, including soil conservation officers and landscape architects, have tended to be master plans which encourage the notion of a final state for the landscape and farm. It might be noted that the final state for everything is death.

In strategic planning, the emphasis is on processes of development which are on-going and respond to changing circumstances. It recognises that complex systems can never be completely described, predicted or controlled but that forces can be identified and worked with to develop a more balanced and productive system. Most importantly, strategy planning can help pinpoint the initial step to get the desired processes moving without later having to undo what has already been done. (1994, p. 21)

The living, evolving system which we call permaculture can only come about as a result of the continuous interaction between the client as designer/practitioner and the elements of climate, soil, plants, animals, buildings and people (p. iii) Melliodora: Hepburn Permaculture Gardens (1995, republished as ebook 2005)

Ben Falk in The Resilient Farm and Homestead

Early in his design process and site establishment chapter, Ben shares a simple diagram showing the planning and design process as an endless cyclic interplay between analysis (see, observe, study), interpretation (consider, decide, affect, apply, mimic) and action (disturb, construct, implement, manage). The diagram blurb reads:

site planning should be continuously fed by a never-ending process of analysing, interpreting, and acting.

In a 2013 podcast with interviewer Scott Man, Ben said:

More and more every year I rely on the planning process as identifying general steps and starting points and trying to visualise … an idealisation of what a place might be in 10 -20 – 50 years, but really using the planning process to identify starting points and letting those starting points then organically drive the actions following those starting points.

If this is not a generating process, I don’t know what is!⁸

Note: See also this more recent podcast in which Ben and I probe these issues directly and deeply.

Toby Hemenway

Although in his book The Permaculture City Toby recommends a fabricating approach, I was struck with this statement which to me is as, if not more, consistent with a generating process:

The point of any design is to move toward some desired outcome-a productive garden, a rewarding business-with as much certainty as possible, some sureness that we’re taking the right steps. … The design process, then, is a program for articulating that purpose and for giving us a sure set of procedures for choosing steps toward it (pp. 25-26)

Jascha Rohr & Sonja Hörster (The Field-Process Model)

It would be an unforgivable oversight not to pay respects here to Jascha Rohr & Sonja Hörster who in this important article not only make the distinction between fabricating (or what they call procedures) and generating processes but explore the characteristics of generating processes in much more detail than I have got to here. Hopefully some day the rest of permaculture will catch up to these exciting thinkers and design practitioners!

Summary

Here is the upshot. In terms of a sound approach to permaculture design process capable of reliably achieving the adapted, nature-mimicking systems permaculture aspires toward, winging it and fabricating get a big thumbs down.

The hybrid and generating approaches get a thumbs up (or big red tick, as the case may be).

In future discussions about permaculture design process, I would love to start seeing the hybrid and generating approaches at the very least being offered as equally viable approaches to permaculture design process.

It is my firm conclusion, however, that the hybrid and generating approaches are not only more viable. They simply are viable, whereas the fabricating approach is unviable as far as reliably realising permaculture’s promise in the world.

I sincerely hope that this effort contributes, even if it is a tiny little nudge, toward a stronger permaculture.

Postscript

I would love to hear what you make of all this, either as a comment below, as an email through the contact form, or, even better, as a guest post which I invite anyone to submit.

References

Alexander, Christopher. The Nature of Order: An Essay on the Art of Building and the Nature of the Universe: Book Two: The Process of Creating Life. Vol. 2. of 4 vols. The Center for Environmental Structure, 2002.
Alexander, Christopher. The Nature of Order: An Essay on the Art of Building and the Nature of the Universe: Book Three: A Vision of a Living World. Vol. 3. of 4 vols. The Center for Environmental Structure, 2005.
Falk, Ben. The Resilient Farm & Homestead, 2013.
Hemenway, Toby. The Permaculture City. Chelsea Green, 2015.
Holmgren, David. Trees on the Treeless Plains: Revegetation Manual for the Volcanic Landscapes of Central Victoria. Holmgren Design Services, 1994.
Holmgren, David. Melliodora: Hepburn Permaculture Gardens. Holmgren Design Services, 1995.
Mollison, Bill. Permaculture: A Designer’s Manual. Tagari, 1988.

Acknowledgements

Thanks to James Andrews and Alexander Olsson for their feedback on a draft of this post.

Endnotes

1. Widespread both outside and inside permaculture
2. See here, here and here.
3. If we agree with Alexander’s proposal that “the world should be created by adaptive processes which act as nature does, itself,” then we have some work to do!
4. Check out this podcast on the matter with Alex Bayley or her posts about agile permaculture here.
5. Consider also this telling comment from renowned ecological design practitioner Dave Jacke: “In reality, I design the overall pattern, implement key pieces after designing them, then redesign as more parts of the system get implemented. I have never had a client where I could implement all at once as a grand expedition! It’s always been piecemeal implementation with design along the way, responding to changes in goals, site and emergent reality as the design goes into place. But having a big picture view, that is, an overall site design to at least a schematic level, is critical to help one work out where to begin the implementation. Then I would design the relevant patches, including their site prep and implementation strategies, and then proceed on the ground. Staking out is a critical part of the process!  Field testing the design in reality, essentially (from a personal email communication received January 28, 2017)”
6. As described of the action-centric design approach here, “no universal sequence of stages is apparent – analysis, design and implementation are contemporary and inextricably linked”
7. He also has claimed that “In broad terms, a generated structure is something that has a certain deep complexity and is created in some way that appears to be almost biological, and reaches deeper levels of subtle structure than we commonly associate with ‘design,’ or with designed objects. (2002, p. 180)”
8. Ben also writes of a design that”The important part to remember is that it’s a guide for next decisions… And they have one more primary purpose: to avoid huge mistakes–for instance, not putting the house in the wrong place or putting the orchard where a road for the eventual barn will need to go…”