The hearsay rule excludes certain out-of-court statements because the person who made an assertion cannot be cross-examined about perception, memory, narration, or sincerity. Courts increasingly confront records produced by machines: a breath analyzer prints a concentration, a server logs a timestamp, or a navigation system reports a location. Some courts declare such outputs nonhearsay because a machine is not a person and therefore cannot make a "statement." The classification is formally neat but analytically incomplete.

Purely automatic measurement does remove sincerity and memory from the usual list of concerns. A thermometer does not lie about what it believes. Yet a machine output can embed human assertions at several layers. A technician selects calibration values; a programmer defines which event counts as an error; an operator may enter the identifier that links a measurement to a defendant. Calling the final record machine-generated can conceal these upstream assertions rather than eliminate them.

The right response is component analysis. A court should identify the proposition for which the record is offered and ask which parts of the production chain assert that proposition. If an instrument directly measures mass according to a disclosed physical process, reliability may be addressed through authentication and evidence of malfunction rather than hearsay doctrine. If the offered proposition depends on a technician's undocumented judgment that a calibration passed, that judgment is a human assertion and should receive ordinary hearsay analysis.

This approach avoids two symmetrical errors. Treating every output as hearsay anthropomorphizes instruments and demands cross-examination where no declarant exists. Treating no output as hearsay allows litigants to bury human testimony inside an automated pipeline. The label on the final page should not decide. The evidentiary question follows the assertion through the system.

Audit trails make the component inquiry more workable. A system that records raw sensor values, later transformations, manual overrides, and identity entries lets a court locate the source of the proposition in dispute. Absence of such records does not automatically make the output hearsay, but it may prevent authentication or obscure a human assertion that requires a witness. The proponent should not gain an evidentiary advantage from designing an opaque pipeline. Conversely, demanding testimony from every programmer who wrote irrelevant background code would recreate the categorical excess the approach is meant to avoid.

The same record may consequently receive different treatment when offered for different propositions. A location log offered to prove the sensor's coordinates may be automatic; the same log offered to prove that a named defendant carried the sensor may depend on a human identity entry. Component analysis explains that difference without declaring the physical file hearsay or nonhearsay for every purpose.