• Member
• September 23, 2026

Generating New Perspectives

Shailendra Jaiswal

A morphological methodology for perspective generation, and an invitation to use it

This document sets out a methodology for generating new perspectives on a problem — deliberately, in a way that can be taught, repeated, and checked, rather than left to intuition alone. It emerged from an extended working conversation applying a Srijan Sanchar framework, AXIM, reflexively to the question of perspective generation itself, and from tracing that work against the established literature it turns out to sit inside. It is offered here in that spirit: as something built openly on real prior work, tested honestly against its gaps, and worth others taking further.

1. Lineage

The methodology does not start from nothing, and it is more useful for saying so plainly. Two traditions converge in it.

Source	Contribution
Zwicky, 1948	General morphological analysis: decompose a problem into independent parameters, then systematically recombine their values into a searchable space of configurations.
Ritchey, 1998–2011 (General Morphological Analysis / Cross-Consistency Assessment)	Formalized pairwise compatibility screening of a morphological box; applied at scale to futures studies and scenario laboratories.
Linstone, late 1970s onward; Mitroff & Linstone, 1993 ("The Unbounded Mind")	The Multiple Perspective approach: Technical, Organizational, and Personal lenses on a single problem, applied to real cases including the Exxon Valdez spill and Bhopal. A fixed three-perspective set, analysis without a formal recombination step.
Gordon, 1961 (Synectics)	Facilitated small-group technique generating solutions via structured direct, personal, symbolic, and fantasy analogies, transferred back to the original problem.
Jaiswal, 2013 ("Multiple Perspective Through Multiple Analogy", MPMA)	A specified source-intermediate-target procedure for group analogical transfer, run at real scale (over 500 Railways managers, roughly 2,000 students) with named, checkable outputs.
Jaiswal and collaborators, 2026 (Probabilistic MPMA)	A mathematical reformulation of MPMA using axis extraction, compatibility tensors, and entropy-based novelty scoring — procedurally ambitious, not yet demonstrated on a worked case.
Srijan Sanchar, AXIM (Morphological Axis Methodology)	A five-route protocol (Ontological Decomposition, Degree-of-Freedom Analysis, Constraint Taxonomy, Stakeholder Value Mapping, Phase-State Analysis) for deriving and testing independent axes of a problem-specific design space.
Srijan Sanchar, RISE (Recursive Invariant Synthesis for Essence)	A validation layer — Find What Survives, Test Through Breakdown, the Regeneration Test, stability under perturbation — for checking whether a candidate structure actually holds, rather than only asserting it.

The central move of this work is bringing these two traditions together: applying morphological analysis's problem-specific, tested axis derivation — historically pointed at product configurations and future scenarios — to Linstone's target domain, generating interpretive perspectives on a present problem, which his own three-lens method never formalized into a combinatorial recombination step.

2. The Promise

Linstone's Technical–Organizational–Personal lenses and de Bono's Six Thinking Hats both work by offering a fixed, small set of perspectives, the same regardless of the problem in front of you. That is their strength and their limit at once: reliable, teachable, but not shaped by what the specific problem actually contains.

The promise of this methodology is a perspective set that is derived from the problem rather than imposed on it — as many perspectives as the problem's own structure supports, no more and no fewer, each one independently tested rather than assumed distinct, and the set as a whole checked for completeness before being used. Where morphological analysis has mostly generated configurations of things or possible futures, this points the same disciplined machine at something more immediately useful in daily reasoning: alternative, defensible ways of seeing what is already in front of you.

3. Concept and Structure

The methodology runs in five stages, each doing real work rather than serving as a formality.

3.1 Problem ontology mapping

Before generating any perspective, the problem itself is decomposed across six categories: its substance (what object or artifact is involved), its process (what happens or changes), its relations (how parts and agents interact), its agents (who acts or decides), its states (what conditions it can be in), and its context (the environment that makes some framings available and others invisible). This mapping is what keeps the axes that follow tied to the actual problem, rather than to a generic checklist.

3.2 Five-route candidate generation

Candidate perspective-generating axes are drawn from five independent routes: Ontological Decomposition (structural and functional breakdowns), Degree-of-Freedom Analysis (the independent variables a designer of the problem could actually set), Constraint Taxonomy (the physical, logical, social, and resource boundaries the problem operates within), Stakeholder Value Mapping (where different parties' values genuinely diverge), and Phase-State Analysis (how the problem's own lifecycle and transitions behave). Each route tends to surface different axes; where two routes independently converge on the same one, that axis is treated as more reliable than an axis only one route found.

3.3 Testing, not accepting

Every candidate axis is tested before being kept. Independence: does varying this axis actually produce a different outcome than varying another? Non-redundancy: is this really a distinct dimension, or the same one under a different name? Actionability: is this a genuine choice available to someone designing a perspective, or a downstream consequence that should be treated as context instead? In practice this step does real pruning — in the worked run of this methodology against the general space of perspective-generation techniques, roughly a dozen initial candidates collapsed and were reclassified down to eleven surviving, independently tested axes, spanning five axis types (configurational, parametric, modal, relational, normative):

Axis	Type	Range
Source multiplicity	Configurational	Single-domain — multi-domain
Transformation operator	Configurational	Positional / rotational / inversional / scalar / spectral / temporal / relational / …
Representational medium	Configurational	Linguistic / formal-mathematical / sensory-instrumental / institutional / cultural
Structural fidelity	Parametric	Metaphorical — formally isomorphic
Directionality	Modal	Permanent departure — round-trip return
Tether strength	Parametric	Loose — tightly bound to target constraints
Protocol intensity	Modal	Lightweight heuristic — structured multi-session protocol
Novelty–actionability orientation	Normative	Surprisal-optimized — yield-optimized
Generation cardinality	Relational	Solo — group-facilitated
Iteration structure	Modal	Single-pass — cyclical/recursive
Observer position	Relational (single-route)	External/analytic — reflexive/participatory

3.4 Completeness checks

Two tests confirm the axis set is doing real work. A coverage test maps known, independently recognizable techniques or framings onto the space to check they land at genuinely different coordinates — if they collapse onto the same point, the axes are not discriminating enough. A surprise test deliberately samples unusual corners of the space — combinations of axis values not obviously drawn from an existing example — to check whether they correspond to real, occupied, or at least plausible configurations, rather than nonsense.

3.5 Generation by sampling

Once the axis set is tested, generating new perspectives becomes a matter of sampling specific corners of the space — deliberately choosing a combination of axis values and working out what perspective that combination actually describes — rather than free-associating. This is where a genuinely useful finding emerged from the work: several of the axis-discovery routes themselves, run in reverse, function as generation operators rather than analysis tools. Constraint Taxonomy run forward finds constraints; run backward, it forces a concrete answer to "what does this look like with one specific constraint removed." Stakeholder Value Mapping run forward finds where values diverge; run backward, it forces the whole problem to be redesigned around one specific stakeholder's values alone. This reversal is not currently described in any of the source literature reviewed and may be a genuine addition in its own right, worth testing further rather than assumed.

4. Possibilities

●       Problem framing and reframing, where a fixed three- or six-perspective checklist is too coarse for the problem's actual structure, and a bespoke, tested set of perspectives is worth the extra derivation effort.

●       Creative problem solving and structured ideation, where the axis-reversal generation operators offer a repeatable alternative to unstructured brainstorming.

●       Organizational strategy and scenario-adjacent work, extending Ritchey's own established use of morphological analysis in futures studies from generating multiple possible futures to generating multiple present interpretations of the same situation.

●       Teaching and facilitation, since the method's stages — decompose, generate, test, sample — are explicit enough to be run by a group without requiring an expert's tacit judgment at every step, in the spirit of Linstone's and Gordon's own facilitated formats.

5. Challenges

These are stated plainly because a methodology asking to be adopted by others should be judged on its actual limits, not a flattering account of itself.

●       Nothing generated through this methodology so far has been run through a validation layer like RISE's Find What Survives, Test Through Breakdown, or Regeneration Test. Every perspective produced to date is an untested candidate, not a demonstrated one.

●       A direct honest audit of one batch of ten sampled perspectives found that four were reconstructions of already-named techniques (pre-mortem analysis, autoethnography, TRIZ-style reversal, the extreme-user method), one was a known weakness in experimental design rather than a genuine technique, and only three to four were assessed as real departures from existing practice. The yield of genuinely new perspectives per sampling run should be expected to be modest, not comprehensive.

●       No worked application of this specific methodology exists yet at the scale or with the rigor of Jaiswal's 2013 MPMA trial — several hundred participants, a real institutional problem, named and checkable outputs. Everything produced so far has been reasoned through by a small number of people in a single extended session.

●       The axis-reversal generation operators are a promising finding but were noticed incidentally rather than deliberately designed and tested; only four of the five routes have been tried in reverse, and the fifth (Ontological Decomposition, reversed) remains untried.

●       Some candidate axes remain unresolved rather than cleanly settled — for instance, whether two axes describing when and how a perspective shift is triggered are truly independent or are the same dimension viewed from two directions.

6. Road Ahead

●       Run RISE's validation steps against a real batch of generated perspectives, to separate genuine restructurings from elaborations that only feel novel.

●       Run the methodology once, properly, against a real problem with a real group — the way Jaiswal ran MPMA against Railways challenges and Linstone ran the Multiple Perspective approach against the Exxon Valdez spill — and compare the perspectives it produces against what an unstructured session produces on the same problem.

●       Test the fifth route, Ontological Decomposition, in reverse, to complete the picture of which axis-discovery routes double as generation operators and which do not.

●       Position the methodology explicitly against its own lineage in any future write-up — Zwicky and Ritchey's morphological analysis, Linstone and Mitroff's Multiple Perspective approach, Gordon's Synectics — rather than presenting it as free-standing, since its actual contribution is the specific combination of these, not an unrelated invention.

●       Build a light, repeatable template — a problem ontology map, a route-by-route candidate log, an orthogonality test record, and an axis card format — so the methodology can be picked up and run by someone who was not part of its original development.

This document sets out a method, not a finished result. Its honest state, as of this writing, is a tested structure with promising early findings and no validated output yet — worth using, and worth checking rigorously as it is used.