When the goalposts have moved too far

How judicial power has become a matter of definition, not demonstration

When I was a boy I used to play soccer in the school playground and my local park. It was always an informal affair, often with two bags acting as goalposts at each end. The park had a slight slope in places. The boundaries of the pitch were sometimes nebulous. But it was recognisably the same game as official football.

Everyone understood what counted as a goal. Everyone knew what it meant to win.

But what would happen if the goalposts at one end were widened to the full width of the pitch, while at the other were narrowed to less than the width of the ball? We would still have players, a ball, the pitch, goalposts, kicks, even “goals”. On paper, it might look like the same game.

But is it?

At some point, something essential has been lost. The structure that made the game meaningful — that made scoring possible and contest real — has been altered beyond recognition. What remains is no longer football in any meaningful sense, even if we continue to call it that.

That question matters more than it first appears.

Because something similar has been happening elsewhere — not on a playing field, but within the operation of the English magistrates’ courts. The rules have been adjusted in subtle but consequential ways. The same vocabulary is used: “court”, “determination”, “justice”. The same result is produced: a conviction.

But when you look closely, the underlying structure has shifted.

---

Over the last week I have been using multiple AI systems to test those “rules of the game” adversarially — to ask, again and again, what it actually means for “a magistrates’ court to determine a charge”. What emerged was a pattern:

Every time you move to challenge the system — to say “this is unlawful” — the structure adjusts.

The state’s goalposts narrow; yours widen. According to the rules, everything remains valid. The referee — at the High Court — sees no foul.

But is it still the same game it claims to be?

What this reveals is something deeper than procedural unfairness. It points to a form of collapse — not of outcomes, but of meaning. A system in which abstraction, presumption, and inference absorb every challenge. Words retain their form but lose their substance.

The historic requirement to show that power has actually been exercised in a given case begins to disappear. What remains is a system that declares itself valid — and treats that declaration as sufficient.

And it has one critical dependency.

The “sport” only holds if we accept that this is still the same game
— even when it plainly isn’t.

Readers who have been following my “ghost court” saga and the recent Part 8 claim will know I have been trying to answer a deceptively simple question:

Does a court operating under the high-volume, paper-based Single Justice Procedure still require a visible act — administrative or judicial — by which it becomes vested of a case?

Or is it sufficient that the statutory machinery as a whole operates, with the status of judicial power effectively conferred on the outcome itself?

As it turned out, I ran into an invisible procedural wall. The issue was one of orthodox statutory construction, framed in entirely conventional terms. But I did not have the right vehicle. Part 8 allows pure questions of law to be determined, but not across the civil-criminal boundary. To reach the issue properly, the claim needs to be anchored within an actual criminal case.

That does not mean the work was wasted. Quite the opposite: it exposed the underlying abstraction model more clearly — and, more importantly, where its limits might lie.

In the meantime, I found myself in a curious interval. The court had struck out my claim, but the order had not yet been sealed. During that period, unaware the case was over, I continued to refine the argument. If anything, the procedural dead end sharpened the question:

What is the strongest possible defence of this abstraction model?

To test that, I set up a structured exchange between two AI systems. One (ChatGPT) generated increasingly focused challenges to the State’s position. The other (Grok) defended it — consistently, and without concession. Each time the challenge appeared to reach bedrock, I pushed again — over 15 successive rounds.

What emerged from that process was what I came to think of as a “squirm path.” Not as an accusation of bad faith, but as a description of what happens when a position is forced to adapt under sustained pressure. If nothing else, it shows how challenging the status quo was essentially impossible before this technology arrived.

What happened was revealing, if subtle. At each stage, the bureaucratic system preserved its legality — but only by loosening something else in the structure: the need for identification, for attribution, for a discernible act, and ultimately for a meaningful distinction between a decision (“you are found guilty”) and its outcome (“pay this fine”).

Tracing each step in that evolution would take an article in itself. What matters here is the end point — the final configuration of the field of play, and where the goalposts have ultimately come to rest.

---

What emerged from the squirm path was not a contradiction, but a stable configuration.

At its endpoint, the system no longer depends on any identifiable judicial act in the individual case. No requirement exists to identify a judge, a tribunal instance, or a moment at which the case was decided. Those features — ordinarily taken as the substance of judicial power — need not appear in the record, nor be reconstructable from it.

Instead, the system operates on a different basis.

The statute defines a set of procedural conditions:

• the issuing of a written charge,

• service of the Single Justice Procedure Notice,

• the absence of a request for a hearing, and

• the recording of an outcome.

Once those conditions are satisfied, the resulting conviction or acquittal is treated as the determination of “a magistrates’ court”.

The determination is not evidence of an act.
It becomes the act, as (re-)defined by statute.

At that point, the distinction between an exercise of judicial power (“I think this person is guilty”) and the outcome of the process (“computer says convicted”) disappears. The two are legally identical. There is no separate event of adjudication that can be identified apart from the result itself.

The notional court produces judicial determinations in law, but it is externally indistinguishable from a robot having done the work.

In the case record, the process is structurally identical to an administrative decision such as a parking ticket, tax assessment, or welfare entitlement: a procedurally completed file yields a binding outcome with no observable act of adjudication beyond the result itself.

The same logic applies to the existence of the tribunal. The case record does not identify which justices acted, and does not provide any case-specific basis from which that tribunal could be reconstructed. Identification operates only at the institutional level — “a magistrates’ court” — not at the level of the individual case — “this magistrates’ court”.

Verification follows the same pattern. There is no independent means, from the case record, of confirming that a tribunal in fact exercised jurisdiction over the charge. What exists is the recorded outcome, together with the statutory rule that classifies that outcome as the determination of the court. The system treats that classification as sufficient.

In this model, jurisdiction itself is constituted in the same way. The fact that “a magistrates’ court has determined the charge” is not established by reference to an identifiable act or tribunal in the case. It is created by the statute’s definition: once the procedural conditions are met and the outcome is recorded, that fact is taken to exist.

There is therefore no underlying requirement for a case-level reality — no discernible act, no particular tribunal, no identifiable moment of decision — corresponding to the exercise of judicial power. The jurisdictional fact, and the determination itself, arise because the statute defines them to arise under those conditions.

The result is a model in which judicial authority over the individual is constituted by definition and institutional assertion, rather than by any independently identifiable exercise of power in the case.

What sustains that model is not evidence, but a form of perpetual presumption:

The assumption that the required judicial act has taken place, even where no such act is visible or reconstructable in the case itself.

Expressed in its simplest form:

1. A set of procedural conditions is satisfied.

2. An outcome is produced in the administrative system.

3. The statute (re-)classifies that outcome as the determination of a court.

And that classification is sufficient to constitute the exercise of judicial power.

---

Once the magistrates’ court system is understood in this way, a further consequence follows.

If the determination of the charge is legally identical to the outcome itself — and if that outcome is constituted by statutory definition — then there is no longer any independent act, precondition, or decision event within the case to which a challenge can meaningfully attach.

In a conventional model, challenges operate by reference to something identifiable:

• a tribunal that can be shown not to exist or not to have been properly constituted

• a decision that can be examined for error

• an act that can be said not to have occurred

• a precondition that can be shown not to have been satisfied

Each assumes that there is a distinguishable element within the case — something that can be isolated, tested, and, if necessary, disproved.

In this model, those elements are no longer independently present.

The tribunal is not identifiable at the level of the case.
The act of determination is not distinct from the outcome.
The exercise of jurisdiction is constituted by definition.
The relevant conditions are satisfied, not by demonstration within the record, but by the system’s own classification of the result.

There is nothing left in the case to get hold of.

The outcome is deemed valid because the system defines a valid outcome as being what the system produced. The space in which voidness would ordinarily be argued is reduced or removed.

The effect is not that challenge is formally excluded. The routes remain: appeal, rehearing, application, review. But the structure to which those routes would ordinarily attach has changed. I faced this myself: which court am I challenging, what jurisdiction, which case? Nothing can be pinned down into concrete form.

So a challenge can still be made, both formally and procedurally. But it no longer operates on an identifiable act or decision within the case. Instead, it operates against a defined state — a result which the law treats as already constituting the exercise of judicial power.

This has a practical consequence.

If the outcome is, by definition, the determination of the court, then showing that no identifiable act occurred, or that no tribunal can be reconstructed from the record, does not, in itself, disturb that determination.

Those are not defects within the model. They are features of it.

The space in which challenge traditionally operates — the gap between what is said to have occurred and what can be shown to have occurred — is narrowed. Not by the removal of rights in formal terms, but by the absence of anything within the individual case against which those rights can take effect.

It would be self-certifying authority — except the definitions have been altered so far that it is deemed not to be.

This is exactly what I experienced in my Judicial Review. “Which court is said to have convicted me?” — the question is treated as meaningless.

Because to the system, it now is.

---

The difficulty with the model described above is not simply that it is unfamiliar. It is that it departs from the way legal authority is ordinarily structured and expressed.

Most statutes — particularly those governing courts — are written in terms of acts. They describe things being done:

• a court hears a case

• a judge considers evidence

• a tribunal determines a charge

These are verbs. They point to events: something happens, at a particular point in time, performed by an identifiable body. The language assumes that legal consequences follow from those acts.

In the model reached here, that relationship is altered. The grammar remains that of action, but the substance has become classification.

While the statutory language remains the same, but the function of those verbs shifts. “Determination” no longer refers to an identifiable act within the case. It refers to a state defined by the completion of a process. The word is preserved, but its referent changes; the computer database entry substitutes for the thing it represented.

This has several effects.

First, it removes the need for a trigger.

In a conventional structure, legal consequences attach when something happens — when a court determines, when a judge orders, when a tribunal decides. Those acts act as triggers: they mark the point at which authority is exercised.

Where the outcome itself is treated as the determination, that trigger is no longer distinct. The consequence does not follow from an act; it is constituted by definition. There is no separate moment at which authority can be said to arise.

Second, it alters the role of abstraction.

Abstraction in law is not unusual. Courts and statutes often describe institutions in general terms as a convenient shorthand. But ordinarily that abstraction resolves into something concrete in the individual case — a particular bench, a particular decision, a particular act that can be pointed to.

Here, the abstraction does not resolve. “A magistrates’ court” remains at the level of definition, without requiring any identifiable instance in the case. The concept floats as an idea, untethered to anything that can be pointed to.

Third, it affects determinacy.

Legal systems depend on the ability to say not just that something has occurred, but precisely what has occurred. That requires a degree of specificity: a decision, made by someone, at some point, on some basis.

Where those elements are not required to be identifiable, the determination becomes less determinate in that sense. It exists as a legal fact, but not as a clearly bounded event.

There is an irony in how those positing the abuse of “strawmen” and legal fictions are dismissed as “sovereign citizens” or “freemen of the land”, while the state expands the use of legal fictions in criminal cases, outflanking traditional rule-of-law limits.

Fourth, it reduces legibility.

For the individual subject to the decision, the ability to understand how authority has been exercised depends on being able to follow the chain from process to act to consequence. If the act is not distinct from the outcome, that chain shortens. What remains is the result, together with the bare assertion that the necessary steps have been taken — in effect, “trust us”.

This is the problem I have encountered with what I call “ghost courts”: a proliferation of labels that do not resolve to a single, identifiable tribunal in law. The system is not merely illegible; it is anti-legible — the labels point to an act-based model of attribution, while the statute operates on a scheme-based one.

In my own case record, the adjudicating body appears under multiple identities — “North Cumbria Magistrates’ Court”, “North and West Cumbria Magistrates’ Court (1752)”, “Carlisle Magistrates’ Court”, and simply “the court” — with no mechanism on the face of the documents to reconcile them into a single tribunal.

Finally, it changes the relationship between description and reality.

If the statute defines the outcome as the determination of a court, and no further case-level verification is required, then the system does not depend on demonstrating that a particular tribunal acted in a particular case. It depends on the correctness of the classification — reflected in a computer entry, not a court order.

We can still say things about the individual case and its validity, because the statutory framework still constrains what outcomes can be produced procedurally. But it remains true that, at the level of the individual case record, the label “determined by a magistrates’ court” does not correspond to an independently identifiable act.

It is a classification applied by the system itself — judicial power is declared to exist without any identifiable judicial actor or act in the case record. In other words, courts can produce judicial power without any identifiable act ever having taken place; the authority exists as a classification, but is empty of any underlying exercise of power in the case itself.

Taken together, these features mark a shift in structure. Not in the existence of legal authority, but in how that authority is expressed and recognised:

• From acts to outcomes.

• From identifiable events to defined states.

• From something that can be pointed to, to something that is asserted to be so.

---

Once the structure is seen in these terms, the issue is no longer one of procedure, but of definition.

At what point does something cease to be an exercise of judicial power, and become instead a result that is treated as if such power had been exercised?

The statutory scheme continues to describe the outcome as a determination of a magistrates’ court. But if that determination does not require any identifiable act, tribunal, or moment of decision in the individual case, the question arises whether the label is doing passive descriptive work, or active classificatory work.

Is this still a genuine judicial act, or a state which the law declares to be equivalent to one — a mere simulation?

That leads to a further question: is there a minimum constitutional content to the idea of a court determining a case?

Article 6 of the European Convention of Human Rights speaks of a tribunal established by law. The common law speaks in terms of courts acting, deciding, determining. Those concepts have historically implied an exercise of judgment that is, at least in principle, capable of being identified and attributed.

If the legislative model does not require that in the individual case, then where is the floor?

How much indeterminacy can be introduced before the concept itself is affected? Is it sufficient that the system, taken as a whole, guarantees that decisions are made somewhere, by someone? Or must it be possible to say, in a given case, what was done, by whom, and when?

A related question concerns verification.

If the existence of the adjudicative act is not visible on the face of the determination, then verification is necessarily deferred. It becomes something that may be established later, if challenged, through internal records or disclosure.

But how far can that deferral be taken?

In practice, verification is not costless. It requires time, knowledge, and often formal proceedings — appeal, application, or judicial review. Those routes carry their own risks and thresholds. For many cases, they are not realistically pursued — as I found in my case, where none of Case Stated appeal, Section 142 review, and Judicial Review determined lawfulness in my motoring case.

To what extent can the system rely on the possibility of later verification, rather than contemporaneous legibility, without altering the nature of the authority it exercises? As noted above, my experience suggests that, in practise, there is no realistic path to challenge the system — to check whether it has minimum operating conditions, and whether they were met in your own case.

There is also a question of language.

Statutes governing courts are framed in verbs: courts determine, judges decide, tribunals hear and resolve. These verbs ordinarily point to acts. If, in operation, those acts are instead treated as states — defined by the completion of a process and the recording of an outcome — then the language is doing something different from what it appears to do.

How far can that shift go before the connection between the words and what they describe becomes attenuated? Should statutes, especially in criminal law, be read using ordinary English? Or is it acceptable for a whole parallel set of meanings to exist to support abstraction?

None of these questions depend on asserting that the system is unlawful. They arise from taking the statutory scheme seriously on its own terms, and following its logic to its endpoint.

They are questions about where the boundary lies.

---

What this analysis reveals is not a breakdown of legality, but a shift in what legality itself depends on. The system continues to operate. The rules are applied. Outcomes are produced and enforced. In that formal sense, nothing has failed.

But the basis on which those outcomes are recognised as lawful has changed.

Where legal authority was once grounded in acts that could be identified, attributed, and, if necessary, verified, it now rests increasingly on re-definition and institutional assertion. The requirement that authority be legible in the individual case — that it be possible to see what was done, by whom, and how — is reduced.

This lowers epistemic legitimacy in order to preserve formal legality.

There may be reasons for that shift: efficiency, scale, the demands of high-volume justice. But the effect is cumulative. Each adjustment preserves the structure of the system while loosening the connection between its language and what can be observed within it.

There comes a point at which that process reaches a limit.

When the goalposts have moved far enough, the question is no longer whether the rules are being followed, but whether the activity being carried out is still the same game those rules were designed to govern.

The system can continue to describe it as such, and enforce its outcomes on that basis. But doing so depends on a further step — that those subject to it accept that description as true, even where the underlying structure is no longer recognisable in the individual case.

That is where the difficulty lies.

Because the issue, at that point, is no longer purely legal.

Within the legal system, these questions are difficult to surface. They tend to arise only at the highest appellate levels, where the structure of the system itself becomes the subject of scrutiny. At lower levels, the machinery operates on its own assumptions, and the questions do not easily take hold.

Outside that frame, however, the question becomes simpler, though no less serious.

What counts as authority?

Is it enough that a system defines its outcomes as the exercise of judicial power? Or does authority require something more — something that can be observed, in the individual case, as an act rather than an assertion?

Those are not questions that can be resolved by procedure alone.

They are questions about recognition: moral, political, and, ultimately, personal.

Whether we accept an authority that cannot be identified, traced, or verified in the case it binds is not only a matter of law.

It is a matter of what we, the public, are prepared to recognise as real.