Here's a startlingly modern way to understand your own mind, taught a thousand years ago in Kashmīr. Your thought-constructs — vikalpas, the stories you hold about reality — are "like programs downloaded from the Internet that affect how your computer operates."1 You downloaded them "from parents, teachers, television, and so on, from before the moment we can understand language onward." Some you're aware of; "others manage to download themselves under our radar." They installed on the hard drive of your brain and "can even rewrite the code of our operating systems." Download a virus — a deeply false, disempowering story — and "it can take over completely, forming a limited view that seems impossible to escape."1
Vikalpa-saṃskāra is the practice of consciously reprogramming this software — the core of what Abhinava Gupta called the śākta-upāya, the Empowered Means.1 It begins the moment you realize that "the way we see and experience reality is largely dictated by the network of programs that have been installed," and that there are "entirely different ways of experiencing reality unavailable to us until we do some reprogramming."1 Studying and internalizing spiritual philosophy is, Wallis says, literally "brain-washing" — "cleansing ourselves of programs not in alignment with ultimate Reality" — and doing it consciously is an expression of your svātantrya-śakti, your power of autonomy: "now you are consciously directing the writing of the code instead of automatically installing whatever is presented to you, as you did as a child."1
The crucial move that makes this work — and keeps it from being naïve "positive thinking" — is a reframe of what a good vikalpa even is. Abhinava distinguishes śuddha-vikalpa (thought-constructs aligned with reality) from aśuddha-vikalpa (those not aligned).2 But here's the catch: "no vikalpa can express reality directly" — nothing articulable in words is completely or permanently true.2 So you're freed from "needless debate about which propositions are 'true' and which 'false'" and can instead focus on which are effective "in bringing about direct experience of reality and which are not."2
This is the key difference between the two types. Non-aligned vikalpas "generate an endless series of further thought-constructs, a sterile snake biting its own tail." Aligned ones "naturally lead you to (and then dissolve in) the wordless, storyless direct experience of reality: totally open, sharply real, and absolutely free."2 So a śuddha-vikalpa isn't a true thought (no thought is fully true); it's a thought shaped to self-destruct — one that points beyond itself, brings you to the edge of direct experience, and then dissolves, "its work done."3 The test of a good story isn't whether it's accurate but whether it deletes itself into reality rather than spawning more stories.
The mechanism Abhinava gives is precise and answers the obvious skeptical objection. A postmodernist asks: how do you know the new "truth" program is more aligned with reality than the old one — haven't you just swapped a story that made you miserable for one that makes you happier, with no guarantee of accuracy?4
Abhinava's refutation: "the experience of reality grounded in differential mental constructs can be undermined, destabilized, and eventually obliterated through the cultivation of thought-forms that are purified, refined, and aligned — but not the other way around."5 The true experience of reality, once reached, "cannot be undermined or destabilized by any dualistic thought-construct." Why? Because "the experience of Reality-as-it-is is not contingent upon thought-constructs."5 So what does the śuddha-vikalpa actually do? It doesn't create the experience of reality. It "serve[s] to destroy the aśuddha-vikalpas that are the substratum of the false experience of duality." Once those are destroyed, "the purified thought-forms naturally dissolve, their work done, and we simply see things as they truly are."5 Reality "is eternally self-manifest and needs no vikalpa to manifest it. The source of all light does not need to be illumined; it is self-luminous."5
So the engine works like a cleaning agent that evaporates after it's dissolved the grime: the śuddha-vikalpa attacks and destroys the false stories that were obscuring self-luminous reality, and then — having no further job, since reality was never produced by any thought — it disappears, leaving direct seeing. (The sequence: aśuddha-vikalpa → śuddha-vikalpa → nirvikalpa, the thought-free direct perception.)3
And this gives a brutal test for whether you've reached reality or merely a refined story: "If someone else's opinion or argument or words in any form can rob you of the feeling or power or integrity of your spiritual experience, even a little bit, then your so-called 'experience' either is or has become a vikalpa, not yet rooted all the way into Reality."6 This is "why gurus with wisdom will send their disciples back home to their families... or tell them to get a job" — around people who think you're a freak for practicing, "you will see your real attainment."6 A vikalpa can be shaken by words; reality can't.
The most practical and revealing part is how to compose your own śuddha-vikalpa — and why the obvious version doesn't work.
Start by identifying a negative self-message, an aśuddha-vikalpa you run. Wallis's was "I'm so stupid," which he used to punish himself with when he made a mistake — until a teacher pointed out: "That's a good way to ensure you keep repeating the behavior."7 The reasoning is sharp: "the ego's priority is to maintain its self-image. It would rather be right than happy" — so if you convince yourself you're stupid, you'll keep doing "stupid" things to confirm the self-image, giving the ego the satisfaction of being right.7 "Because we believe the self-image... we see the world in such a way that confirms our view."7
Now the trap. The obvious antidote to "I'm so stupid" is "I'm actually really smart!" — and Wallis says flatly that "such a feeble rejoinder will not be successful in displacing an impure thought-form that has taken deep root and plagued you for years."8 Why does the positive affirmation fail? Because, per Abhinava, the new purified thought-form must do two things: it "must be very powerful" and it "must terminate in the ultimate Reality."8 A flat positive affirmation ("I'm smart," "I'm a good mother") does neither — it stays at the level of the personal ego, trading one self-image for another, and the ego can argue with it. The only way to make a vikalpa powerful enough and self-transcending enough is to "take it all the way to God."8
So instead of "I'm actually really smart," Wallis offers: "The divine intelligence that created this whole universe dwells within me as me, and by contacting it, I can understand anything I truly need to know in this life."8 That is a śuddha-vikalpa — powerful, and it "can take me beyond itself" (it points past the personal self to the divine, so it can dissolve into direct experience rather than just installing a nicer self-image). The diagnostic that you've found a real one: "you will probably be uncomfortable repeating it. You might feel silly saying it, or you might start to cry."9 Wallis's student, working on "I'm a bad mother," tried "I'm a good mother" (feeble), then "I am a manifestation of the Divine Mother" — and "burst into tears," then said "But I don't really believe it," so they modified it to "I am a manifestation of the Divine Mother in the process of realizing and embodying my true nature" — true enough to hold, challenging enough to work.9 The case study's lesson: rewriting a deep negative belief fails when you stay at the level of the ego swapping self-images (the ego just argues back), and works only when the new thought roots in something bigger than the ego — which is also exactly what lets it dissolve rather than become a new attachment.
Vikalpa-saṃskāra has a defined procedure (the bhāvanā-krama, stages of contemplation), and it runs on a threefold support — sat-tarka (sound reasoning), sad-āgama (true scripture), and sadguru (a true teacher) — the tripod Abhinava says gives the process stability.10 To work a teaching-passage: (1) find a passage that resonates (adjust the wording until it does); (2) understand it literally, learn the words, traditionally memorize it (the vaikharī level); (3) ponder it deeply, asking "how would I experience the world if I felt the truth of this fully?" (madhyamā); (4) when you start to get bored with it, sit and meditate on it as a mantra, invoking grace, letting what arises arise (paśyantī); (5) for days, sleep with it, wake with it, examine your resistance to it; (6) then let it go and watch how the internalized wisdom shows up in life.11 You know a teaching is assimilated when "you can explain it clearly using language wholly different from that in which you learned it" — "if you can't show it, you don't really know it."11 To compose your own antidote: identify a negative self-message, then build a replacement that is powerful and terminates in God (not a flat ego-affirmation), authentic enough that you can hold it without lying to yourself — and test it at the mirror, repeating it "no matter how silly you feel, until you get past the point of feeling silly."9 Throughout, keep the deeper aim in view: you're not installing a permanent better-story, you're installing a story built to dissolve into direct seeing.
The practice comes from Abhinava's Tantrāloka and Tantrasāra (chapter 4), with Wallis's commentary and personal examples.2 Wallis himself raises the strongest objection (the postmodern one: maybe you've just swapped a sad program for a happy one, with no claim to accuracy) and gives Abhinava's refutation — but the refutation has a soft spot Wallis half-concedes.4 The objector's deeper version: any sufficiently deeply-rooted program (at the paśyantī level) shapes experience "without our conscious awareness," so how do you know the nirvikalpa "direct experience" isn't itself just the deepest, most invisible vikalpa? Wallis's first reply is almost a shrug — "even if the objection is correct, we could do worse than to impose... a view that makes us joyous and fills us with love for all living beings."4 Open question: is Abhinava's claim that direct experience "can never be supplanted" by any thought-construct a genuine epistemological proof of having reached unconstructed reality, or is "unshakeable by words" compatible with being an extremely deep construct — and does the brutal test (can words rob you of it?) actually distinguish reality from a sufficiently-rooted program, given that the objection is precisely that the deepest programs are the ones we can't see shaking?
This page stages Abhinava directly against a postmodern skeptic, with Wallis as both relayer and participant — and the encounter is unusually fair because Wallis voices the skeptic's best shot. Abhinava's position: aligned thought-forms can dismantle false ones and then dissolve, revealing self-luminous reality that no thought can touch. The skeptic's position: you've just installed a different program; "reality as it is" might be your deepest, most invisible programming. They converge on the mechanism of how belief shapes experience (both agree deeply-rooted constructs shape perception below awareness — that's the whole premise). They diverge on whether there's a bottom: Abhinava says there's a self-luminous reality beneath all constructs that the cleaning-agent thought reveals; the skeptic says it's constructs all the way down. Where Wallis sides with Abhinava, he leans on phenomenology — those who've had nirvikalpa perception "will never doubt its reality, since it is an entirely different order of experience."12 But that's an appeal the skeptic can't check from outside, and Wallis knows it, which is why his fallback ("even if the objection is right, a loving joyous view beats a miserable one") quietly shifts the ground from truth to pragmatics. Here is the honest limit of the practice's self-justification: vikalpa-saṃskāra demonstrably works (it changes experience, dismantles suffering-stories), but whether what it reveals is unconstructed reality or just the most refined and unshakeable construct is a question the practice can answer experientially (the practitioner becomes certain) but not argumentatively (the certainty is exactly what the skeptic doubts) — and Wallis's willingness to fall back on "it's worth doing either way" is the tell that he sees the gap.
The bridge: this is a page about deliberately replacing maladaptive beliefs with better-aligned ones — but with the distinctive twist that the replacement is engineered to self-destruct into direct experience rather than be permanently maintained, which both connects to and sharpens cognitive-change methods elsewhere.
The Sharpest Implication If the goal of refining your beliefs is to install ones that delete themselves, then the entire self-improvement project of "thinking better thoughts" is, at best, a transitional stage — and getting attached to your good new beliefs is just a more comfortable cage than the bad old ones. The radical part is that a śuddha-vikalpa you cling to has become an aśuddha-vikalpa: the very thought that was supposed to dissolve into reality, if held onto, becomes one more story standing between you and reality. Take that seriously and the test turns merciless and personal — "if anyone's words can rob you of your experience, it's still just a thought" — which means the spiritual identity you've built, the beliefs you defend most fiercely, the experiences you're proudest of, are exactly the candidates for being refined constructs rather than reality, because you'd defend them. The deepest practice isn't acquiring the right beliefs; it's using beliefs to dismantle beliefs and then having the nerve to let even the good ones evaporate — which is why a wise teacher sends you home to people who think you're a freak, to find out what survives when the supporting stories are stripped away.
Generative Questions