The words an NPPF case turns on

These are the phrases the National Planning Policy Framework (2026) uses without defining. Leaving them open is deliberate: it gives a decision-maker room to judge. But it also means the outcome of a planning case often depends on language nobody has pinned down, which is worth understanding before you rely on any of it.

Where a phrase appears in two places with different reach, that is noted. There is at least one instance where the same three words mean materially different things depending on whether your land is in the Green Belt.

“substantially outweighed”S4, S5, GB7

Undefined, and the whole balance turns on it. Materially higher than the old significantly and demonstrably formulation, with nothing in the Framework saying how much higher.

“settlement”Annex B

The Framework points to development plan definitions. Where the plan draws no boundary, the extent is argued from what is predominantly built-up. Land allocated or permitted that will form part of the built-up area counts too. Villages lying within and defined as part of the Green Belt are excluded entirely.

In plain terms: It means the built-up part of a town or village, and it also includes land already earmarked or permitted for building. A village inside the Green Belt does not count, however built-up it looks.

“evidenced unmet need”S5(1)(j) against GB7(1)(g)(ii)

The same phrase is open in one place and closed in the other, and for small and self-build sites that is the most important asymmetry in the document. S5(1)(j) says including, but not limited to, a five year supply failure: any evidenced need can count. Footnote 41 to GB7(1)(g)(ii) says that for housing it means a five year supply shortfall including any buffer, or a Housing Delivery Test result below 75%: on a strict reading nothing else counts. At appeal, inspectors have counted self-build demand under the closed version too, so far always alongside a five year supply shortfall, so whether it can satisfy the test alone is undecided. Annex A(3) adds one limit on the open version: where the authority has a five year supply and a Housing Delivery Test result above 75%, a standard method figure higher than the adopted requirement is not evidence of unmet need for five years from adoption.

In plain terms: The same three words mean different things in different places. Outside the Green Belt, any evidenced shortage counts. Inside it, only two specific housing supply measures do. That single difference decides a lot of self-build cases. One thing counts nowhere: where the council is up to date, the national formula giving a bigger number than its plan is not a shortage for five years after adoption.

“people wishing to commission or build their own homes”HO1(2)(i)

This is how self-build demand becomes an assessed need, which is what earns the substantial weight at HO7. Where the council's evidence has never assessed it, that gap is itself the argument.

In plain terms: This is how self-build gets counted as a real housing need. If your council has never assessed it, that omission becomes your argument instead.

“physically well-related”S5(1)(h)(ii), S5(1)(j)(i), GB7(1)(h)(ii)

There is no test and no distance. Footnote 28 covers only land beyond an undeveloped allocation. This is the main battleground for edge-of-settlement sites.

In plain terms: Nobody has said how close is close enough, so this is usually what gets argued about on a field at the edge of a village.

“a scale which can be accommodated”S5(1)(h)(iii), S5(1)(j)(i), GB7(1)(h)(iii)

The words existing or proposed let you rely on infrastructure that does not yet exist. How firm the proposal must be is untested.

“limited infilling”S5(1)(e), GB7(1)(c)

Neither limited nor group of houses is defined, and both are read tightly at appeal. The pinch point for multi-plot self-build frontages.

In plain terms: Filling a gap is allowed, but how big a gap and how many houses count as a group are undefined, and inspectors read both narrowly.

“strongly contribute”Annex B grey belt, via GB2

Land is grey belt where it does not strongly contribute to purposes (a), (b) or (d). Those purposes rest on the phrases large built-up area, town and historic town, none of which is defined. Purpose (c), safeguarding the countryside from encroachment, is deliberately left out of the test.

In plain terms: Whether land strongly does one of those three jobs is a judgement, not a fact, and councils and inspectors can legitimately disagree. Protecting open countryside in general is deliberately not one of the tests.

“inappropriate or not possible”L3(2)(c)(i)

The escape from the density floor for smaller schemes has two parts. Being below the major threshold is necessary but not sufficient: you must also show the standard cannot sensibly be met on this site. Preferring bigger plots is unlikely to be enough.

“net developable area”L3(2)(c)

The 35 and 45 dwellings per hectare floors bite on the net developable area. Annex B keeps access roads, hard landscaping, private gardens, ground level parking and incidental open space inside it, and takes out only land for strategic infrastructure, parks and public squares, and open space serving the wider green infrastructure network, including significant biodiversity net gain and drainage features. On a small site that is normally the whole site, which is why this tool measures against the site area.

In plain terms: The minimum is measured over the part you develop, and that includes the roads, gardens and parking, not just the houses. Only things like a public park or a large drainage pond come out, which rarely applies to a small site, so the whole site is usually the right figure.

“disproportionate increase in size”S5(1)(c) against GB7(1)(b)

Same words, different baselines. S5 measures against the building as it stood at publication of the Framework. GB7 measures against the original building, meaning 1 July 1948 or as first built if later.

In plain terms: A real trap. Outside the Green Belt you measure against the building as it is today. Inside it you measure against 1948, so any extensions since then do not raise your starting point.

“original building”GB7(1)(b), footnote 40

The building that existed on 1 July 1948, or if constructed later, as it was built originally.

In plain terms: In the Green Belt, the size you are compared against is the building as first built, or as it stood in 1948. Later additions do not count in your favour.

“truly outstanding”HO11(1)(e)

The exceptional quality design route for an isolated one-off, and the only limb in the housing policies that turns on architecture rather than location.

In plain terms: It means exceptional by national standards, not simply well designed. Very few schemes clear it.

“reasonable prospect of being served”Annex B, well-connected station

Lets planned upgrades or an agreement with the rail operator satisfy the frequency threshold. How firm the prospect must be is undefined.

“substantial harm”HE5(2)(c), HE6(5), HE7(3)

The only steer is HE5(2)(c): substantial harm occurs where the proposal would seriously affect a key element of the asset's significance. Neither key nor seriously is defined, and this line decides whether the HE6(4) balance or the HE6(5) refusal applies. HE5(3) adds that what counts is the effect on significance, not the scale of the development, so a single dwelling can cross it.

In plain terms: This is the line between harm the council weighs up and harm that means a refusal. All the rules say is that it happens where the scheme would seriously affect something central to why the place matters. A small scheme can still cross it.

“clear and convincing justification”HE4(2)

Any harm to a designated asset's significance needs this, including harm from development within its setting. The phrase is undefined, but it demands more than simply giving reasons, and it is a separate requirement from the HE6(4) public benefit balance: winning that balance does not supply the justification.

In plain terms: Any harm to something formally protected needs a clear and convincing reason. Nobody has defined how convincing, but it is more than just explaining yourself.

“wholly exceptional”N6(2), HE6(6)

The bar for losing irreplaceable habitat such as ancient woodland or a veteran tree, where N6(2) also requires a suitable compensation strategy, and for substantial harm to or loss of heritage assets of the highest significance under HE6(6). Higher again than exceptional circumstances.

“exceptional circumstances”S5(4)

The fallback for anything outside the S5 list. The benefits must substantially outweigh the adverse effects, which expressly include harm to the character of the countryside and to sustainable patterns of movement.

In plain terms: Where no route fits, this is what is left, and it is a high bar. The harm counted against you expressly includes harm to the character of the countryside and the extra car journeys.

“very limited weight”Annex A(2)

What a development plan policy materially inconsistent with the national decision-making policies now gets, unless the plan was examined and adopted or made against this Framework. Not zero: how little is for the decision-maker, case by case. A policy is not downgraded simply for predating the Framework; the inconsistency has to be shown. Most restrictive rural housing policies adopted under earlier editions sit here.

In plain terms: Where an old local policy clashes with the new national rules it now counts for very little, unless the plan was checked against the new rules when it was adopted. That is often what unlocks a rural site.

Looking for hope value, option agreements or Section 106? Those are in the land development definitions.

Reflects the National Planning Policy Framework (2026), published 17 August 2026. This page last reviewed 21 August 2026. Read the framework.