Cosmology · What cosmology asks and how we know

Established findings and open questions

Build a map of confidence: distinguish the sky's measured records, the cosmic history they support, the parameters inferred within a model, and physical questions still being tested.

Begin here with curiosity. No previous astronomy course or advanced mathematics is required.

Back to the chapter

After reading, you can

Classify a cosmic statement by its evidence and model dependence.
Separate evidence for a dark component's effects from identification of its physical nature.
Explain why a tension or a successful fit alone does not establish a discovery or a final theory.

A map of confidence is more useful than a list of mysteries

You can know that a shadow is present before knowing which object casts it. The analogy is limited: cosmology uses several quantitative measurements and physical models, rather than a single shadow. But it captures a useful logical distinction. Evidence for an effect and identification of its underlying cause are different achievements.

Cosmology has established observational records, well-supported explanatory frameworks, conditional parameter estimates and open mechanisms. These categories are connected, not a ranking from valuable facts to worthless guesses. A precise model-dependent result can be highly informative when its assumptions are explicit. An attractive proposal with little discriminating evidence should remain a proposal.

Read uncertainty at the level of the statement. A disagreement about one inferred parameter does not make a measured background radiation field disappear. Conversely, the existence of that field does not prove every feature of an early-Universe theory. Keeping the claims separate allows a model to be improved without discarding the observations that made it useful.

What the sky records

We measure extragalactic spectra, brightnesses, positions and shapes; we detect a microwave background with a near-thermal average spectrum and small spatial variations. Each record requires calibration and processing. The observations are nevertheless reproducible targets that an adequate account of cosmic history must address.

Joint interpretation of distance–redshift patterns supports large-scale expansion. Microwave radiation and suitable primordial light-element abundance measurements support an earlier hot, dense state. The distribution and gravitational effects of matter constrain how structure formed. These conclusions do not depend on one spectacular image or one quoted age.

The background photons mostly last scattered when the early plasma became sufficiently neutral for radiation to travel relatively freely. That statement differs from saying the background was created at recombination. The first follows the physical history of radiation already present; the second would misidentify the origin of its photons.

Do not confuse well-supported with literally assumption-free. Even a wavelength measurement uses calibration. The important question is whether alternative choices consistent with established physics substantially change the conclusion, and whether independent checks constrain those choices.

Precise numbers come with a model label

A cosmological parameter summarizes a feature of an adopted model. A fit can infer a cosmic age, matter density or present expansion rate by comparing predicted observables with data. The uncertainty is conditional on the analysis model, likelihood and other assumptions; it is not a guarantee that every unmodelled effect has been ruled out.

Base ΛCDM uses a cosmological constant, cold dark matter and a specified family of initial fluctuations, with spatial flatness fixed. Λ is the Greek capital lambda, used here for the cosmological constant. The name is shorthand for a framework with mathematical predictions, not a complete microscopic theory of every ingredient.

The Planck 2018 parameter analysis, published in 2020 and corrected in its archived version, found strong agreement with base ΛCDM across temperature, polarization and lensing analyses. Quoting its parameter values requires that dataset and model context. The numbers are a documented result, not timeless constants independent of how they were inferred.

Two studies can infer different values because they use different observations, calibrations, priors or models. A meaningful comparison first aligns definitions and uncertainty conventions and identifies shared information. Only then can the size and interpretation of a discrepancy be assessed.

Different claims need different evidence labels
Statement	Appropriate label	What would be overstated
An instrument detects a calibrated microwave spectrum	Measurement with processing and uncertainty	Calling it a direct detector reading of cosmic age
A fitted expansion history yields a cosmic age	Model-dependent inference	Quoting it without the model or dataset
Several gravitational observations support a nonluminous component	Evidence within a physical framework	Claiming its particle species has thereby been identified
An early accelerated phase could explain primordial patterns	Tested theoretical proposal	Presenting one mechanism as directly observed

Effects can be constrained before mechanisms are known

Dark matter is used to explain several gravitational phenomena, including structure and gravitational effects inconsistent with visible matter alone in the standard framework. Its microscopic identity remains open. Detecting a new particle would require evidence that connects its properties and abundance to those cosmic effects. An unexplained detector event is not automatically the cosmological dark matter.

Cosmic acceleration is inferred through observations interpreted with expansion models. In base ΛCDM a cosmological constant accounts for that behaviour. Broader dark-energy models allow other dynamics. Establishing acceleration does not by itself show which microscopic explanation is correct, and fitting one present-day parameter does not determine every possible future.

Comparing modified-gravity proposals with dark-component models requires the complete evidence: background expansion, gravitational lensing, structure growth, local tests and internal physical consistency as applicable. An alternative is not successful merely because it addresses one anomaly, and an unknown mechanism does not make every proposed alternative equally supported.

These open questions are productive because proposed explanations can differ in observable consequences. A candidate might predict a particle interaction, a change in structure growth or a particular time dependence of expansion. The next step is to identify such a consequence and test it with controlled uncertainties.

The earliest past and farthest future require extra care

Inflation proposes an early period of accelerated expansion and supplies mechanisms that can generate primordial fluctuations. Measurements constrain features of those fluctuations and eliminate some models. Such agreement does not identify a unique inflation mechanism, directly reveal an inflaton field, or establish an absolute beginning.

Classical equations can lead to an initial singularity when extrapolated backward. That signals a limit of that description in an extreme regime; it does not, without further justified physics, describe creation from nothing. Proposed earlier phases and quantum-origin scenarios need their own evidence labels and tests.

Possible futures depend on assumptions about dark energy, gravity, particle stability and long-term astrophysical processes. Continued expansion in one model is a conditional prediction. Black-hole evaporation is a theoretical quantum-field result, while proton decay is a hypothesis constrained by searches in particular channels. Later chapters give these possibilities their own homes rather than presenting one speculative timeline as an observed destiny.

A tension is a problem to investigate

A tension is a mismatch between results expected to agree under specified assumptions. Before naming new physics, researchers test calibration, selection, analysis, shared data and the treatment of uncertainties. A mismatch can expose a problem in the model, in a measurement pipeline, or in the comparison itself.

Statistical significance is calculated within an uncertainty model. It does not identify a cause, and it can change when systematic effects or a different sample are included. To assess a discovery claim, ask which alternative was tested, how the test was selected, whether the result persists in independent data and what physical explanation predicts it.

The frontier will change as new surveys and experiments report results. This lesson provides a way to read them rather than a claim that one snapshot settles every current dispute. The dedicated inference, tensions and dark-energy chapters must attach dates, datasets, models and uncertainties to particular results.

A useful closing question is: what would make me revise this statement? If the answer is a clearer calibration, a different probe, a physical consistency check or a new predicted signal, the question has become a research path. Wonder is most durable when it stays connected to that path.

Worked reasoning

Repair three overconfident headlines

How would you revise: 'The CMB directly measures the Universe's age'; 'Dark matter's gravitational effects identify its particle'; and 'A parameter disagreement proves new physics'?

Separate the radiation measurement from the age calculated within a cosmological fit.
Separate evidence for a gravitational component from the additional detection needed to establish its microscopic identity.
Separate the existence of a discrepancy under stated assumptions from an explanation of its cause.

Use: 'CMB data constrain age within a stated model'; 'Several observations constrain dark matter's effects while its identity remains open'; and 'The discrepancy motivates checks and model tests.'

Try the reasoning

If one model fits several datasets well, has every physical ingredient been identified microscopically?
Show the answer and reasoning
No. Predictive success and microscopic identification are different claims.
The model can correctly represent a component's effects while leaving its detailed nature unresolved. Additional evidence is needed to identify a mechanism.
Why should a far-future scenario state assumptions about dark energy and particle stability?
Show the answer and reasoning
Those assumptions can change the predicted expansion and the survival of material systems.
A conditional calculation is useful only when readers can see what conditions make it applicable. A future cannot be treated as observed.
Does disagreement between two inferred expansion rates invalidate all evidence for a hot early Universe?
Show the answer and reasoning
No. The disputed inference and the broader evidence are different claims.
The disagreement may motivate revisions, but a proposed revision must still account for the measured radiation, abundances, expansion and structure evidence it aims to explain.

What to carry forward

Separate measured records, explanatory frameworks, conditional parameters and open mechanisms.
Evidence for effects does not automatically identify their microscopic cause.
A successful fit is a testable achievement; a tension is a specific unresolved comparison.
Keep ultimate-origin and far-future claims conditional on the physics they require.

References

Sources supporting the explanations, physical distinctions and examples in this lesson.