What you are allowed to put when you do not know
The answer is not no. Two named instruments already contain worked patterns for a stated unknown and both require it bounded and quantified, which means you may be able to say what you do not know if you can put a number on it. A third instrument runs the other way, and there an unbounded unknown is a reason the product may not be placed on the market at all.
On this page
The short answer, and the condition attached to it
The answer is not no. Named instruments in EU law already carry worked patterns for saying, on a label and in a regulated document, that part of something is not known, and the pattern they use is consistent: the unknown is stated as a proportion rather than as a shrug. This page is about what you may defensibly publish when you do not know. How we establish whether anything is known in the first place, and the vocabulary this page borrows for it, is on How we know.
The condition is the whole of it. Every permission described below is a permission to publish a bounded, quantified unknown inside the instrument that grants it. None of them is a permission to leave a field empty, and none of them travels to a field in another regime. A business that can say what proportion of a product it cannot account for is in a position the law has written procedures for. A business that can only say it does not know is in a position at least one instrument treats as a reason not to sell.
That is the shape of the answer. What follows is where each pattern comes from, how far it was read and where it stops.
Before this page is the right page
The chemicals regulation's supplier information duties are both capped at the information available to the supplier, so a supplier who genuinely does not hold something can lawfully say so, and the gap then arrives at your desk as an empty field. Nothing in that article gives a brand a right to demand information from its own supplier. If you have not yet exhausted the asking, getting evidence out of suppliers is the earlier page and it is worth finishing first.
Two instruments that already permit a stated unknown
Both are ordinary, long standing, unglamorous instruments. Neither was drafted with product passports in mind. What makes them worth reading is that somebody has already had to solve this problem in a place where getting it wrong is a labelling offence, and the solution they reached is the same solution twice.
Classification, labelling and packaging
The provision read for this page handles a mixture whose components are not all characterised. The mixture is classified on the components whose acute toxicity is known, and the label and the safety data sheet carry a stated percentage of the components of unknown acute toxicity. The unknown is not hidden, it is not guessed at and it is not left blank. It is quantified and printed next to the classification it qualifies.
That permission does not transfer to another instrument or to another field. It belongs to that instrument and to the documents that instrument governs, and a percentage of unknown acute toxicity on a safety data sheet says nothing about what may be written in a product passport attribute. We read one annex provision, not the act.
Textile fibre labelling
The labelling regulation carries two provisions of the same family. Other fibres are permitted up to stated proportions, so a composition can be complete without naming everything in it. And a composition may be given as unspecified where it is hard to state at the time of manufacture, which is a named legal category for a value the manufacturer cannot pin down.
Neither of those permissions transfers to another instrument or to another field. Both live inside a labelling instrument, on a label, for the fibre composition of a textile product, and neither is a licence to write unspecified into a passport attribute that happens to sound similar. The relevant provisions were read; the annexes were partly truncated on some retrieval routes and nothing here rests on them.
Two instruments, drafted decades apart for unrelated purposes, reached the same answer. Say how much you do not know, in the same place you say what you do.
The distinguishing factor is boundedness, not sector
It is tempting to read the two patterns above as chemicals law being permissive and textiles law being permissive, and to conclude that some sectors tolerate uncertainty and others do not. That reading does not survive the counter case in the next section, which comes from a third sector and runs the other way.
The line that actually separates them is whether the unknown has an edge. A stated percentage of components of unknown acute toxicity is an unknown with an edge: a reader knows exactly how much of the mixture the classification does not cover. Other fibres up to stated proportions is an unknown with an edge for the same reason. In both cases the business has done real work, because you cannot state a proportion you have not measured or bounded, and the statement is checkable by somebody else.
An unbounded unknown has none of those properties. It cannot be checked, it cannot be compared against the next declaration and it gives a reader no way to tell a small gap from a total one. That is why it is the version the law is hardest on, and it is also why the practical work in front of most businesses is not finding the missing value. It is establishing how much of the product the missing value covers.
This is the point at which the estate's own vocabulary becomes useful to a business rather than to us. The four resolution states and the five kinds of absence set out on How we know exist to force exactly that question before anything is written down, and they are worth borrowing whether or not you use any of our tooling. A value that is not applicable, a value nobody has looked for and a value that was looked for and not found are three different rows, and only one of them is a candidate for a bounded statement.
The counter case, where not knowing stops the sale
A page that showed only the permissive patterns would be misleading in the direction that costs money, so the counter case sits here at the same weight rather than in a footnote. The deforestation regulation does not offer a bounded unknown at all.
Three of its provisions were read. The route to market runs through a statement carrying information that is adequately conclusive and verifiable, and where that information is not held, the product may not be placed on the market. There is no proportion to state, no unspecified category and no permitted default. The instrument does not ask how much of the product is unaccounted for, because the answer does not change the outcome.
Two consequences follow, and they point in opposite directions for the same business. The first is that a permission found in one instrument tells you nothing about a second: the classification pattern and the labelling pattern do not transfer here, and reasoning from them into this regime would produce a product that cannot lawfully be sold. The second is that where an instrument works this way, the money is spent upstream. There is no publishable fallback to fall back on, so the only thing that resolves the field is the information itself.
What this page does not say about that instrument
The estate watches the deforestation regulation and does not publish it as a subject. Nothing here describes which products it reaches, when its requirements bite or what its due diligence process consists of. It appears on this page for one structural reason: it is the worked example of a regime where an unbounded unknown is a bar on the sale rather than a qualification on the label.
What a properly constructed fallback looks like
There is a third shape worth knowing, and it comes from a regime with no connection to textiles whatsoever. The carbon border implementing regulation is the clearest example in EU law of a fallback that a drafter designed on purpose rather than a gap somebody tolerated.
Two of its provisions were read. Three features do the work together.
- Published default values, time limited. The fallback figure is supplied by the regulator rather than invented by the declarant, and the permission to use it has an end.
- A cap on how much may be estimated. In complex goods there is a stated ceiling on the share of the figure that may rest on estimation, so the fallback cannot quietly become the whole answer.
- A duty to state the methodology used. Whatever route produced the number travels with the number, which is what makes it checkable rather than merely permitted.
None of that transfers. It is a different regime, a different obligation and a different reader, and nothing in it permits an estimate in a textile field or in a passport attribute. Its value here is as a template for recognising the real thing: when somebody tells you that estimation is allowed, the three questions are who published the default, what caps it and whether the method has to be disclosed. A permission missing all three is usually not a permission at all.
What the passport framework says, and what it does not say
Two findings, and the second one needs stating carefully because it is the sentence most likely to be misquoted off this page.
The first is that the framework's accuracy standard is unqualified. The data is to be accurate, complete and up to date. There is no tolerance written beside it, no materiality threshold and no wording that softens it for a value that is hard to obtain.
The second is a finding about the text rather than about the law. A term search of the consolidated framework text returns no occurrences of estimate, estimated, estimation, default value, proxy or average value. That is not a prohibition on estimation and it is not a permission either. It is the observation that the framework act contains no estimation vocabulary to reason from, which is a different and much smaller statement. The search was run on 28 August 2026 and is recorded at medium to high confidence with an instruction to re-run it before publication, and it will be re-run before this page ships.
Two limits on how far that finding reaches. Articles nine to fifteen of the framework have not been read verbatim by this estate, so this is a term search across a consolidated rendering rather than a close reading of the passport provisions, and the finding could change under one. And the framework works by empowerment: what a textile product actually has to carry is a matter for a delegated act, and that act does not exist yet, so nobody can say whether any of the patterns on this page will be adopted for a textile field. That is genuinely unknown rather than undecided by us, and it is recorded as unknown rather than filled with the nearest neighbouring rule.
What a schema can carry when the value is absent
Knowing what an instrument permits is only half of it. The other half is whether the thing you are typing into can express the distinction at all, and this is where the subject usually collapses, because a form field with a value or a blank cannot hold the difference between a question that does not arise and a question nobody answered.
Below is a survey of the carriers this page actually read, with how far each was read. It is a survey of what these schemas can express when a value is absent. It is not a survey of what any business does, and it deliberately does not cover commercial product data schemas or marketplace form specifications, because this estate has not read them and a row we cannot support is worse than a missing row.
| Carrier | What it can express when the value is absent | Can the unknown be bounded | Read status |
|---|---|---|---|
| Classification label and safety data sheet | A stated percentage of components of unknown acute toxicity, printed beside a classification made on the known components. | Yes, and the boundedness is required rather than optional. | One annex provision read at the consolidated text. |
| Textile fibre label | Other fibres up to stated proportions, and an unspecified composition where it is hard to state at the time of manufacture. | Yes, as a proportion of fibre mass. | Relevant provisions read. Annexes partly truncated on some routes. |
| Carbon border report | A published default value in place of the actual figure, with the methodology used stated alongside it. | Yes, and capped for complex goods. | Two provisions read. |
| Deforestation due diligence statement | Nothing. Either the information is adequately conclusive and verifiable or the product may not be placed on the market. | No. There is no partial state to record. | Three provisions read. |
| Substances information under the chemicals article | Whatever the supplier holds. Both limbs are capped at the information available to the supplier. | Not by the article itself. The cap sits on the supplier rather than on a stated proportion. | The article and the relevant definition read at the consolidated text. |
| The passport framework's own data requirements | Not established by us. The accuracy standard is unqualified and the text carries no estimation vocabulary to build an absence state from. | Not established. The per product act that would settle it does not exist. | Recitals and some provisions read. Articles nine to fifteen not read verbatim. |
Two things are visible across those six rows. None of the six distinguishes between a value that does not apply, a value that is not known and a value that is known but withheld, so a business carrying all three states internally has nowhere in these carriers to put the difference. And one of the six has no absence state at all: the product either qualifies or it does not go on the market, which is a design decision rather than an oversight.
A field can also be structurally impossible to fill rather than merely difficult, which is a different finding again, and it rests on evidence about the interface rather than about the product. Where two sources give you different values instead of none, the procedure is different again and it is on when two sources disagree.
The decision tree, six gates
The gates run in order and each one is answered for one field, under one instrument. There is no universal answer to any of them, which is why every gate in the drawing carries the instrument that answers it and the phrase in its own scope. A permission read off one row is not a permission on any other row.
The drawing is the structure. What follows is the same six gates as questions you can put to one field in front of you, with the source that answers each one.
May I write something other than the value in this field?
-
1Does the instrument that governs this field say anything at all about what to do when the value is absent?
- YesRead what it permits, and read the scope it permits it in. The scope is the part people skip.
- NoThen nothing has permitted a substitute for that field, and this page cannot supply one.
- Not sureSilence in a text is neither permission nor prohibition. Record the field as unanswered rather than as a yes.
SourceEcodesign for Sustainable Products Regulation (EU) 2024/1781, consolidated textChecked 28 August 2026
-
2Does what it permits require you to put a number on the part you do not know?
- YesThat is the bounded pattern, and it is the one that gets published. A stated percentage of components of unknown acute toxicity is the worked shape of it, in that instrument's scope.
- NoYou are relying on a shape you have not found in the text. Go back a step.
SourceRegulation (EC) No 1272/2008 on classification, labelling and packaging, consolidated textChecked 28 August 2026
-
3Can you state the method that produced whatever you are putting in the field?
- YesPublish the method beside the value. Where EU law permits a fallback it tends to require exactly that, along with a cap on how much may be estimated.
- NoAn estimate whose method cannot be stated is a number with nothing behind it, and no instrument read for this page permits one.
SourceCommission Implementing Regulation (EU) 2023/1773 on carbon border adjustment reportingChecked 28 August 2026
-
4Does the instrument make adequately conclusive information a condition of placing the goods on the market?
- YesThen an unbounded unknown does not qualify the sale, it stops it. The gate is market access rather than labelling.
- NoThe unknown is a labelling question rather than a market access question, in that instrument's scope only.
SourceRegulation (EU) 2023/1115 on deforestation-free productsChecked 28 August 2026
It cannot tell you what any other instrument permits. Every answer above is good only inside the instrument that gave it, for the field that instrument governs.
What this page does not say
The subject invites one particular misreading, so the guard is written in the copy rather than left to the reader. Nothing above tells anybody they may leave a named field blank. What it does is set out what named instruments permit, in their own scope, for the documents those instruments govern, and it states in each case that the permission does not travel.
Four more limits, plainly.
- No generalisation between regimes. A mixtures instrument, a fibre labelling instrument, a carbon border instrument and a deforestation instrument are four regimes. A pattern in one is a worked shape to recognise, never a permission to reuse.
- No claim that the passport framework prohibits estimation. The finding is that a term search of its consolidated text returns no estimation vocabulary. Absence of text is neither permission nor prohibition, and the search is to be re-run before this page ships.
- No score for how complete anything is. This estate publishes no completeness measure, on this page or anywhere, because a number that rates a business against its own gaps would turn a method into a product and would be the exact temptation this page exists to refuse.
- No prediction about a textile field. Whether any of these patterns is adopted for a textile attribute is a question for a delegated act that does not exist. It is unknown, it is stated as unknown and no sentence on this page reads forward from it.
What survives all of that is short and useful. Find out how much of the product your gap covers. Where the instrument governing the field has written a bounded pattern, use its words, in its scope, with your method stated. Where it has not, treat the field as unresolved rather than as empty, and keep the record of what you tried, because that record is the only thing that distinguishes a gap somebody worked on from a gap nobody looked at.
You might want to read next
Sources
-
The annex provision on mixtures containing components of unknown acute toxicityCELEX 02008R1272-20250101In forceRelevant provisions reviewed
The first worked pattern for a compliant stated unknown. One annex provision was read: a mixture is classified on the components whose acute toxicity is known, and the label and the safety data sheet carry a stated percentage of the components whose acute toxicity is not. It is a mixtures instrument. Nothing in it reaches a product passport field, and the passage that uses it says so in the same breath.
-
The provisions on other fibres and on unspecified compositionCELEX 02011R1007-20180215In forceRelevant provisions reviewed
The second worked pattern, and the only one on this page that touches textiles at all. Read for two things: that other fibres are permitted up to stated proportions, and that a composition may be given as unspecified where it is hard to state at the time of manufacture. Both are permissions inside a labelling instrument and neither is a permission anywhere else.
-
The provisions on the due diligence statement and on placing on the marketCELEX 32023R1115In forceRelevant provisions reviewed
The counter case, carried at the same weight as the permissive patterns rather than as a caveat. Three provisions read. Without a statement carrying adequately conclusive and verifiable information, the product may not be placed on the market, so an unbounded unknown is a bar rather than a qualification. The estate watches this instrument and does not publish it as a subject, so nothing here describes its scope of application.
-
The accuracy provision, plus a term search of the consolidated textCELEX 02024R1781-20240628In forceRelevant provisions reviewed
Declared for two findings and no more. The accuracy standard the framework sets is unqualified. And a term search of the consolidated text returns no estimation vocabulary, which is a finding about the act rather than a permission or a prohibition. Articles nine to fifteen have not been read verbatim by this estate and the page says so where it uses this source. The term search is to be re-run before publication.
-
The provisions on default values and on reporting the method usedCELEX 32023R1773In forceRelevant provisions reviewed
The worked model of a permitted fallback that does exist in EU law: time limited published default values, a cap on how much of a complex good may be estimated, and a duty to state the methodology used. Two provisions read. A different regime entirely, and it appears here for the shape of the permission rather than for anything it permits a textile business to do.
-
Arts. 3(33) and 33CELEX 02006R1907-20250101In forceRelevant provisions reviewed
Read to establish the cap rather than the duty. Both limbs of the article are limited to the information available to the supplier, which means a supplier's own ignorance can be a lawful answer and it arrives at your desk as your empty field. Nothing in the article gives a brand a right to demand information from its own supplier, and that correction is why this page exists one step downstream of asking.
Help someone else make sense of product passports.