Research and version note This is a version 0.1 research draft in Who Maintains the World? It draws on public material from the BIPM and Australia’s National Measurement Institute current to 2026. It explains an institutional structure and is not calibration or legal-metrology advice for a particular instrument.
Who Maintains the World? · Article 5
Quick read
A kilogram looks like a quantity already written into nature. In practice, an international system translates a definition into comparable measurements. Since 2019, the kilogram has no longer been defined by the mass of one international prototype near Paris. It is established through the fixed numerical value of the Planck constant within the International System of Units. The definition has escaped dependence on a single artefact, but not maintenance. Laboratories still need demanding experiments to realise the unit, compare results, and calibrate lower levels of standards.
Australia’s National Measurement Institute maintains the country’s highest-level standards and provides traceability to the SI through an unbroken chain of calibrations. A supermarket scale, manufacturing instrument, medical device, and law-enforcement instrument have different requirements; they are not all directly compared with a national standard. Confidence moves through levels and also depends on uncertainty statements, certificates, accreditation, and review.
What is maintained is not an unchanged object but the comparability of measurements across places and time within stated uncertainty. Calibration is not a permanent transformation that makes an instrument correct. It compares indication with reference, records conditions, and establishes a relationship from which correction may follow. Instruments drift and environments change, so a certificate has a scope and temporal meaning.
Standards exercise coordinating power. Trade, health, environmental regulation, and legal liability depend on measurement, making methods, laboratory recognition, and dispute processes publicly significant. Participation cannot vote an experimental result into truth. Legitimacy comes primarily from open definitions, reproducible realisation, international comparison, demonstrated competence, and challengeable procedures.
Disagreement is not eliminated by traceability. Two results can differ because their uncertainty, environmental conditions, sampling or method differ, even when both refer to the SI. A competent system preserves enough information to investigate the discrepancy. The word traceable should never function as a ceremonial guarantee detached from the actual chain and uncertainty relevant to the result.
Metrology is a form of distributed maintenance. Nobody personally verifies every kilogram, yet a traceable institutional chain preserves common comparability. Reliability does not come from a centre announcing that a number is correct. It comes from each link identifying what it can measure, its uncertainty, and the boundaries of its responsibility, while remaining open to comparison.
Where does the kilogram on a package come from?
A customer buying goods labelled one kilogram does not ask the shop to explain the SI. A scale displays a value and law and commerce treat it as shared language. The institutional structure becomes noticeable only when scales disagree.
One unit has to cross different uses. Food trade concerns fair quantity. Pharmaceutical and chemical work may need much smaller uncertainty. Manufacturing requires parts produced at separate sites to fit, and science needs other laboratories to reproduce results. If a kilogram were meaningful only on one device, exchange and evidence would contract to the local site.
Measurement consequently has two aspects. The physical world supplies relationships that can be observed. Institutions agree on units, methods of realisation, and evidential chains. Agreement does not make a unit arbitrary; it allows different experiments to refer to one definition. A constant of nature will not calibrate a shop scale by itself.
From the international prototype to the Planck constant
For many years the kilogram was defined by the mass of an international prototype. The artefact offered an intuitive centre but also a fundamental limitation. If the mass relationship between the prototype and its copies moved slightly, it was difficult to say which had changed. A global unit depended on one object, and realisation depended on a comparison chain leading towards it.
The 26th General Conference on Weights and Measures adopted a revision taking effect on 20 May 2019. All seven SI base units are now constructed from fixed defining constants. The BIPM SI Brochure explains that the kilogram is defined by fixing the numerical value of the Planck constant h. The old prototype lost its definitional status.
This change is sometimes described as making the unit eternal. More precisely, the definition is no longer attached to an artefact that can be contaminated, damaged, or lost. Laboratories still have to realise it through accepted methods. A Kibble balance, for example, connects mechanical and electrical quantities through a demanding experiment. Its operation needs instruments, environmental control, software, reference standards, and expertise.
Stable definition and revisable realisation are compatible. As experiments improve, institutions update the practical account of how a unit can be realised without redefining the unit. Change in method preserves the common structure.
Traceability is not error-free copying
Australia’s National Measurement Institute maintains primary national standards and enables Australian measurement to be traced to the SI. Traceability means an unbroken chain of calibrations in which each comparison contributes a statement of uncertainty.
It is not a perfect value copied down a hierarchy. Every measurement has conditions and uncertainty. As the chain approaches use, equipment capability and environment become increasingly relevant. Whether an industrial scale is adequate depends on permissible error, range, use, and legal context, not only on the abstract existence of traceability.
A calibration certificate records a comparison, method, range, and result at that time. It does not promise that the device will never drift or that a user will operate it correctly. An organisation chooses calibration intervals based on risk and observes performance between calibrations. A one-year interval maintained only because it is customary does not by itself inspire confidence.
The practical value of traceability is that disagreement can travel back through evidence. When two laboratories differ, their standards, methods, uncertainty evaluations, environments, and processing can be examined. Without a chain, the parties can only invoke reputation or repeat the same local method.
Who maintains a common unit?
International institutions created under the Metre Convention coordinate the units, including the CGPM, the International Committee for Weights and Measures, and the BIPM. National metrology institutes realise national standards and demonstrate capability through international comparisons. Accreditation bodies, calibration laboratories, instrument service organisations, regulators, and users continue the chain into practice.
This coordination has no all-seeing centre. The BIPM does not inspect every market scale, and a national institute does not calibrate all factory instruments. A hierarchy of competence and evidence does the work. Each participant is responsible not for declaring the entire world correct but for accurately stating what it compared, how, and with what uncertainty.
Repair and succession belong to the chain. Laboratory equipment ages, reference materials have storage requirements, and software updates alter calculations. Staff need to understand historical data and the effect of change. If a technique is known only by one person approaching retirement, the physical standard may remain while national capability weakens.
Capability is therefore tested through more than possession of expensive equipment. Laboratories compare results, document uncertainty budgets, examine unexpected differences and demonstrate that qualified staff can reproduce the work. A technically sophisticated apparatus without maintained procedures, environmental control and succession is not a national measurement capability. The institution has to preserve the means of finding its own deviation.
Legal metrology and scientific calibration
Scientific metrology concerns comparability and uncertainty. Legal metrology governs measurements used in trade, enforcement, and public protection. Australia’s National Measurement Act 1960 and associated instruments establish legal units and controls. A trade scale, some enforcement devices, and other regulated instruments may require pattern approval, verification, or work by licensed servicing organisations.
Law cannot make an instrument physically more accurate. It can say who is authorised to verify it, what error is allowed, which records are kept, and how a dispute proceeds. Consequences differ by use, which is why the same measurement technology can require different procedures.
Technical standards therefore participate in power. A measurement may determine whether a consumer receives less than promised, a company complies, or evidence is accepted in court. Procedures need to control conflicts of interest and allow affected parties to inspect certificates and contest method or process. A challenge cannot overturn a validated result merely through subjective disagreement; it needs a reason related to evidence.
This is also why the consequence of an error should shape the legal process. A disputed package quantity, a measurement used for a large commercial transfer and an instrument used in enforcement do not impose the same harm. Verification intervals, tolerances and review need a defensible relation to use. Uniform units support fairness, but uniform procedure across every instrument would ignore the different risks that law is meant to manage.
Calibration is not a one-off certificate of correctness
Everyday speech treats calibration as adjustment. Strictly, calibration first compares the indications of an instrument with reference values and establishes a relationship. Adjustment may follow as a separate action. The distinction matters because an adjusted instrument requires verification, while some systems apply corrections without changing hardware.
Measurement uncertainty is not permission for arbitrariness. It characterises the dispersion reasonably attributed to a result under defined conditions. A supposedly exact number without uncertainty often conceals the limit that matters. Compatibility between laboratories is assessed with uncertainty, not by asking whether the final decimals are identical.
Maintainers must prevent a certificate from becoming a permanent shield. Shock during transport, temperature, intensity of use, and repair can change an instrument. High-consequence systems use intermediate checks and control information to detect drift before the next formal calibration. Scheduled and local evidence then support each other.
What remains incomparable after units agree?
Uniform units permit quantities to be exchanged but do not ensure that the measured object has the same definition. Two organisations may record waste in kilograms while classifying waste differently. Medical results in one unit can have different implications because of sampling and population.
Metrological maintenance solves only part of a knowledge problem. It preserves quantitative relationships. It does not decide what policy should measure or whether a threshold is fair. A precise scale may enable fair trade and provide precise evidence for an unjust quota.
Sampling creates another boundary. A balance may measure the submitted material with excellent traceability while the sample fails to represent the shipment, patient or environment about which a conclusion is made. The uncertainty of the instrument cannot absorb a defective sampling design. Maintaining comparability therefore includes stating where metrology ends and another evidential responsibility begins.
This limitation protects rather than diminishes metrology. Separating unit, method, object, and uncertainty prevents physical precision from being borrowed to support a conclusion outside its evidential range.
Does international uniformity suppress local difference?
Common units reduce friction in trade and science, but standardisation can be criticised as an extension of central authority. Historically, measurement systems have been entangled with state administration, markets, and colonial power. Traditional measures also carry local practice and culture.
The issue need not become an absolute choice between global units and local knowledge. International trade, medicine, and climate observations require comparability. Local practices may use classifications and experiential measures better suited to a particular activity. Public institutions should explain when SI traceability is required, where local measure remains meaningful, and what conversion loses.
The legitimacy of a common unit rests on coordination needs and open governance, not on a claim that only one way of understanding the world is possible. Standards should change when science and use require, and nations should participate through international institutions.
Provisional judgment: comparability is the maintained object
The continuity of the kilogram no longer depends on an immutable material object. Its definition changed profoundly; techniques of realisation continue to develop, and instruments are repeatedly calibrated and replaced. What continues is a relationship in which measurements across place and time can be traced to a common unit and compared with stated uncertainty.
That relationship needs physical experiment and institutions. International agreement, national capability, accreditation, certificates, repair, and succession work together. A link that invokes authority without exposing evidence weakens the chain.
Metrology also corrects a stereotype within this series. Invisible maintenance is not always low-status manual labour. It can involve highly specialised laboratories, law, and diplomacy. The common issue is not occupational image. It is whether stability depends on correction and whether those responsible have resources, demonstrated competence, and clear accountability.
Distributed quantum methods and new realisations may alter the hierarchy. More laboratories may realise units directly, but comparison and evidence of competence will remain. The current conclusion is therefore limited but durable: constants of nature stabilise the definition, while the mutual recognition of a kilogram around the world remains a maintained public achievement.
Primary sources and further reading
- Bureau International des Poids et Mesures, The International System of Units (SI Brochure), 9th edition, updated 2026.
- BIPM, Resolution 1 of the 26th CGPM, 2018.
- Australian Department of Industry, Science and Resources, Australia’s measurement system.
- Australian Government, National Measurement Act 1960.
- National Measurement Institute, Measurement Standards Unit calibration services.
Series navigation: Who Maintains the World? series overview
Discover more from Geoffrey Chen
Subscribe to get the latest posts sent to your email.