10.23.5 Tensor Change of Basis Error Pattern Boundary

Tensor Change of Basis Error Pattern Boundary is the classification of the recurring mistakes that arise when performing a tensor change of basis, together with the line separating errors that are genuine mathematical faults from apparent anomalies that are actually correct behavior occurring near a legitimate boundary of validity.

Purpose of Classifying Error Patterns

Separating Mistakes From Legitimate Boundary Behavior

Not every unexpected result produced by a change-of-basis calculation is an error. Some apparent anomalies, such as a component growing large near a coordinate singularity, are the correct and expected consequence of approaching a genuine boundary of validity. Cataloguing recurring error patterns alongside the boundary conditions that can mimic them helps distinguish a computational fault from an inherent feature of the transformation.

Recurring Nature of the Mistakes

The mistakes catalogued here recur across many different specific calculations because they stem from a small number of structural misunderstandings about index placement, variance, and the summation convention, rather than from arithmetic slips specific to any one problem.

Index Placement Errors

Swapping Upper and Lower Index Roles

A frequent error places a contravariant index where a covariant index belongs, or the reverse, which silently applies the direct Jacobian factor where the inverse factor was required, or vice versa.

This error is distinguished from a true coordinate boundary because it occurs everywhere the transformation is applied, not only near a special locus, and it is corrected simply by substituting the correctly oriented Jacobian factor.

Mismatched Primed and Unprimed Labels

A related error occurs when an index intended to belong to the new coordinate system is left unprimed, or the reverse, causing the equation to equate quantities from two different coordinate systems as if they were already expressed in the same one.

Summation Convention Violations

An Index Appearing Three Times

Because the summation convention is defined only for an index that appears exactly twice within a term, once as a superscript and once as a subscript, an index appearing a third time in the same term produces an expression with no defined meaning under the convention, rather than a subtle numerical error.

A Free Index Left Unmatched

Another common pattern leaves a free index appearing on one side of an equation without its matching counterpart on the other side, which breaks the requirement that free indices correspond in name and position across an entire equation and typically signals a missing Jacobian factor or a dropped term.

Confusing Genuine Error With Boundary Effects

Component Blow-Up Near a Coordinate Singularity

When a computed component grows without bound as a point approaches the origin of a radial-type coordinate system, this is frequently mistaken for a computational error. In fact it reflects the intrinsic coordinate system boundary at that locus, and the correct response is to switch to a regular coordinate system for evaluation there rather than to search for an arithmetic mistake.

Non-Invertibility Mistaken for a Broken Formula

If a Jacobian determinant vanishes at a particular point and the inverse Jacobian factors appear to be undefined, this is not evidence that the transformation formula was set up incorrectly. It is the expected signature of a transformation boundary, and the correct diagnosis is to check the determinant directly rather than to re-derive the transformation from scratch.

Distinguishing Symptoms

Errors Present Everywhere Versus Errors Localized to a Set

A genuine notational or algebraic error typically produces incorrect results across the entire domain of the calculation, while a true boundary effect is localized to a specific lower-dimensional set, such as a single point, curve, or surface, where the Jacobian determinant vanishes or a coordinate system is intrinsically degenerate.

Errors That Vanish Under Relabeling Versus Persistent Ones

An index-placement or labeling error can often be revealed by systematically relabeling every index in the calculation and checking whether the same discrepancy reappears in the same place; if the discrepancy instead follows a specific geometric locus regardless of how the indices are labeled, the underlying cause is a genuine boundary of validity rather than a notational mistake.

Preventive Practice

Verifying the Reciprocity Identity

Before trusting a set of computed Jacobian factors, checking that the direct and inverse factors satisfy the standard reciprocity identity through the Kronecker delta catches the majority of index-placement errors before they propagate into a larger calculation.

Explicitly Stating the Domain of Validity

Recording the domain over which a transformation is valid, including any known coordinate singularities or non-invertibility loci, before performing the change of basis prevents legitimate boundary behavior from later being misdiagnosed as a computational error.