The property of a module or morphism failing to be flat: tensoring with the module does not preserve exact sequences, so base change can produce nonexact results and torsion or Tor groups appear.

Principle

Flatness is equivalent to exactness of the tensor functor; nonflatness therefore signals hidden relations or torsion that obstruct the passage of exactness through tensor products and cause base-change anomalies.

Demonstration

Consider the short exact sequence 0→Z --×2--> Z → Z/2Z →0 and tensor with Z/2Z; because Z/2Z is not flat over Z, tensoring destroys exactness and Tor_1(Z/2Z,Z/2Z)≠0, demonstrating nonflatness concretely.

Misapplication

Assuming a base change or pullback preserves kernels and cokernels without checking flatness leads to incorrect fiber computations or false dimension statements in algebraic geometry.

Consequence

Nonflatness causes failure of naive base-change theorems, nonconstancy of fiber dimension, and can produce unwanted torsion in families; it complicates descent, deformation, and cohomology calculations.

Reversal

Flatness restores tensor-exactness: flat modules and flat morphisms allow base-change to commute with taking kernels and preserve fiberwise properties, yielding better-behaved families.

Boundary

A notion for modules and morphisms of schemes or rings; nonflatness does not by itself indicate pathology of all constructions and must be examined relative to the exact sequences or base changes under consideration.

Semantic Tension

Between flatness, projectivity and freeness: projective or free implies flat, but flat need not be projective; nonflatness should not be conflated with lack of freeness alone.

Synthesis

Nonflatness is the systematic failure of tensoring to preserve exactness: it reveals torsion or hidden relations that spoil base-change, fiberwise behavior, and many simplifications one expects in flat families.