The Gold Standard of Universal Algebra | Separating Term Operations from Polynomial Operations with Fixed Parameters

A signature specifies operation symbols and their arities. For the additive group of residues modulo 6, use a binary symbol +, a unary symbol − and a constant symbol 0. The carrier is {0,1,2,3,4,5}, and every calculation is interpreted modulo 6. Ordinary integer labels identify the six residue classes.

The expression x+x is a term in this signature. So is x+(−y), and so is 0. The expression x+1 requires care: the carrier contains the element labelled 1, but the signature as stated contains no constant symbol naming it. An element's existence in the carrier does not automatically make its name part of the formal language.

You can still define the function taking x to x+1 by choosing the element 1 as a fixed parameter. That construction is legitimate, but it belongs to a different description. Calling it a polynomial operation records the extra permission. It should not be silently used as evidence about what the original terms alone can express.

This is similar to specifying the ingredients permitted in a recipe, although the mathematical restriction is exact. Knowing that another ingredient exists in the kitchen does not put it on the authorised list. The analogy explains the bookkeeping; the signature and evaluation rules supply the actual mathematics.

Before simplifying an unfamiliar expression, list its symbols. Which are variables? Which are declared operations? Which are designated constants? Which refer to particular carrier elements supplied from outside the term language? A short inventory often resolves the confusion before any algebraic manipulation is needed.

Contents · Next chapter

A term is a formal expression. A term operation is the function obtained by evaluating that expression in a particular algebra. Different terms can produce the same function. Conversely, the same formal term can produce different functions when interpreted in different algebras of the same signature.

In our modulo-6 additive group, x+x and −(x+x+x+x) give the same output for every x, because 2x and −4x differ by 6x. Their written expressions are different, but their induced unary functions agree. At x=2, both give 4; at x=5, both give 4 again.

A complete table is a useful finite check. For x=0,1,2,3,4,5, the function 2x has outputs 0,2,4,0,2,4. If another expression gives exactly this six-entry list, it defines the same unary operation on this carrier. It need not be the same formal expression or agree in a different algebra.

Now interpret the same two terms in the additive group of integers. At x=1, their outputs are 2 and −4, so they differ. Equality of term functions in one algebra is not a universal identity of every algebra with the same symbols. The domain of interpretation must remain attached to the statement.

When reporting a result, distinguish “this expression is a permitted term,” “these expressions evaluate equally here,” and “this equation holds throughout the chosen class.” The three claims require different evidence. The rest of the guide keeps the function-level question visible while respecting its formal-language starting point.

Contents · Previous chapter · Next chapter

Every unary term in the additive-group signature evaluates to kx for some integer k. This can be proved by following the construction of the term. The variable x has coefficient 1. The constant 0 has coefficient 0. Adding two previously constructed expressions adds their coefficients, while negation changes the sign of a coefficient.

Nothing in those construction steps can introduce a nonzero fixed offset. If the subexpressions are kx and lx, their sum is (k+l)x. Negating kx gives (−k)x. Beginning with x and 0 and repeating these operations therefore stays within the stated form.

Conversely, every integer multiple kx can be built using repeated addition, negation and zero. In modulo-6 arithmetic, only the residue of k modulo 6 affects the function. The six possible unary term operations are therefore 0, x, 2x, 3x, 4x and 5x.

These six functions are distinct because evaluating them at x=1 gives the six different residues. This final check matters: listing six expressions does not automatically establish six different operations. On another carrier, several expressions might collapse to the same function, as our earlier 2x and −4x example demonstrated.

The derivation has two directions. Structural induction shows that no other form can appear. Explicit construction shows that every listed form is available. A few sample expressions would illustrate the answer, but only the two-direction argument classifies all unary term operations in this example.

Contents · Previous chapter · Next chapter

Take the two-variable term t(x,z)=x+z. Choose the carrier element 1 for z and keep that choice fixed. The resulting unary function p(x)=t(x,1)=x+1 is a polynomial operation. Its variable input is x; the value 1 is a parameter, not another changing input in the unary function.

The output table is 1,2,3,4,5,0 for inputs 0 through 5. In particular p(0)=1. Every unary term operation classified in Chapter 3 sends 0 to 0, so p cannot be a unary term operation of the original signature. One carefully chosen input proves the distinction.

If you let z vary, you return to a binary term operation. If you fix z=4, you obtain a different unary polynomial operation x↦x+4. Arity and parameter choice should therefore be recorded separately. A display containing two symbols does not by itself tell you how many variable inputs the final function has.

More generally, begin with a term t(x1,…,xn,z1,…,zm), choose elements c1,…,cm of A, and evaluate t with those last inputs fixed. The resulting n-ary function is a polynomial operation of A. Parameters may be repeated or unused; what matters is that the chosen values remain fixed while the n variable inputs vary.

This definition does not grant permission to use arbitrary new operations. You may insert carrier elements into a term built from the existing signature. You may not introduce multiplication, division, comparison or a conditional instruction unless the original operations can express the intended function or the signature has explicitly been expanded.

Contents · Previous chapter · Next chapter

Allow parameters in the modulo-6 additive group. A term in variables and parameters is an integer linear combination of all its inputs. Once the parameter values are fixed, their combined contribution becomes a constant residue c. Every unary polynomial operation therefore has the form p(x)=kx+c modulo 6.

Every such form is attainable. Build kx using the original operations and then add the chosen parameter c. There are six possible residues for k and six for c, giving at most 36 functions. To show there are exactly 36, we must check that different coefficient pairs cannot accidentally define the same function.

Suppose kx+c and lx+d agree for every x. At x=0, we obtain c=d. At x=1, we obtain k+c=l+c, hence k=l modulo 6. Thus the parameter pair is determined by the function, and all 36 choices are distinct.

For a concrete example, p(x)=2x+1 has outputs 1,3,5,1,3,5. The coefficient 2 controls how the output changes with the input, while the parameter 1 shifts the output list. Calling this function “linear” without specifying whether an offset is allowed can conceal the distinction. Here “affine” identifies the form with a possible offset.

The complete count gives a useful scale comparison. There are 6⁶=46,656 arbitrary unary functions from a six-element set to itself. Only 36 are polynomial operations of this particular additive algebra, and only six are term operations. Allowing parameters enlarges expressiveness, but it does not make every possible input-output table available.

Contents · Previous chapter · Next chapter

For a unary polynomial operation in this example, the values at 0 and 1 determine the candidate affine formula. The offset is c=p(0), and the coefficient is k=p(1)−p(0) modulo 6. This observation is useful for checking a proposed table, provided we do not confuse determining a candidate with proving the table belongs to the class.

Suppose a table begins p(0)=4 and p(1)=1. The candidate coefficient is 1−4=−3, which is 3 modulo 6. The candidate function is 3x+4, with output list 4,1,4,1,4,1. All six entries must agree before the supplied table is accepted as this polynomial operation.

Now change only the last supplied entry from 1 to 2. The first two values still suggest 3x+4, but the complete table is no longer affine. Two values uniquely determine an affine candidate; they do not prove that an arbitrary function is affine. The distinction is the same one encountered when fitting a familiar formula to insufficient data.

A quick invariant supplies another check. The differences p(x+1)−p(x), including the wrap from input 5 back to 0, should all equal k modulo 6. For 3x+4, every successive difference is 3. The altered final entry breaks this pattern.

These checks assume the full table is known and exact. A noisy measurement dataset would require a different statistical problem, not a declaration that a physical system is an algebraic polynomial. Our calculation is a finite mathematical verification with specified operations and no measurement uncertainty.

Contents · Previous chapter · Next chapter

A polynomial operation is built from the algebra's operations and fixed parameters. A homomorphism is a map that preserves the declared operations between algebras. These descriptions are not interchangeable. The shifted function p(x)=x+1 is a polynomial operation on the additive group modulo 6, but it is not a group endomorphism.

The constant requirement already fails: p(0)=1, whereas a homomorphism preserving the designated zero must send 0 to 0. The binary requirement fails as well. With x=1 and y=2, p(x+y)=p(3)=4, but p(x)+p(y)=2+3=5 modulo 6.

The unshifted function h(x)=2x does preserve addition: h(x+y)=2x+2y=h(x)+h(y). It also preserves negation and zero. In this abelian-group example, each unary term operation kx is an endomorphism. That is a property of this particular family, not a universal fact about all unary term operations in every algebra.

For a counterexample to that broader claim, consider the group of permutations of three objects under composition. The unary term x·x is a permitted term operation, but squaring is not generally a group homomorphism in a noncommutative group. For two different transpositions a and b, a² and b² are identities, while ab is a three-cycle and (ab)² is not the identity.

The correct question is therefore specific: which preservation equation is required, and does this function satisfy it? A function's origin as a term or polynomial supplies useful structure, but it does not automatically make the function preserve the operation when applied to combined inputs.

Contents · Previous chapter · Next chapter

Although a polynomial operation need not itself be a homomorphism, homomorphisms interact predictably with term evaluation. If h:A→B preserves the basic operations, applying h after a term calculation in A agrees with evaluating the same term in B on the images of all its inputs.

Parameters count among those inputs. Let pA(x)=tA(x,c). Then h(pA(x)) equals tB(h(x),h(c)). The parameter on the B side must be h(c). Reusing an unrelated element with the same printed numeral or name can break the equality.

Use A as the additive group modulo 6, B as the additive group modulo 3, and h as reduction modulo 3. Let pA(x)=2x+4. The transported parameter is h(4)=1, so pB(y)=2y+1 modulo 3. At x=5, pA(5)=2 modulo 6, whose image is 2. On the other route, h(5)=2 and pB(2)=5=2 modulo 3.

If someone instead chooses parameter 2 in B, the second route gives 2·2+2=0 modulo 3, contradicting the first. The failure is not in homomorphism preservation. It is in the mismatched parameter. Keep the source element, its image and the target interpretation on separate lines.

This is a practical way to read a commuting calculation without a diagram. Compare “evaluate, then map” with “map every input and parameter, then evaluate.” If they differ, inspect whether the operations are preserved and whether every fixed value was transported. A hidden parameter is often the missing piece.

Contents · Previous chapter · Next chapter

Every polynomial operation preserves every congruence of the algebra in the coordinatewise sense. If corresponding variable inputs are equivalent, the outputs remain equivalent when the same fixed parameters are used on both sides. The reason is simple: each parameter is equivalent to itself by reflexivity, and term operations preserve congruences.

In the modulo-6 additive group, consider the parity congruence with blocks {0,2,4} and {1,3,5}. The polynomial p(x)=2x+1 sends every input to an odd residue. Thus any two equivalent inputs certainly have equivalent outputs. In fact, this particular function also sends some inequivalent inputs into the same output class.

The function q(x)=x+1 behaves differently: it swaps the parity blocks. Inputs from the same block still land in the same block, so congruence preservation holds. Preserving a relation does not require fixing each class pointwise or even sending each class to itself.

This explains an apparent puzzle. The shift q fails to preserve the group identity and addition as a homomorphism, yet preserves the parity relation and all other congruences. The preservation requirements concern different structures. One compares operations with a map; the other compares related inputs with related outputs.

Avoid reversing the theorem without justification. A function preserving every congruence need not be a polynomial operation. In Chapter 14 we construct a small algebra where the congruence condition is too weak to characterise the available polynomial functions. A necessary condition becomes a reliable classification only when a converse has actually been established.

Contents · Previous chapter · Next chapter

Let B={0,3} inside the additive group modulo 6. It contains zero, is closed under addition modulo 6, and is closed under negation. It is therefore a subalgebra in the original additive-group signature. Every term operation of the original algebra carries inputs from B back into B.

Now use the polynomial p(x)=x+1, whose parameter lies outside B. Its outputs on B are p(0)=1 and p(3)=4, both outside B. The function is still a legitimate polynomial operation of A, but it does not restrict to an operation on B.

The failure does not contradict closure of B under the original operations. To evaluate x+1, we supplied an input that was not in B. Subalgebra closure guarantees outputs remain inside only when all relevant inputs, including any parameters used in a term calculation, lie inside the subset.

Replace the parameter 1 by 3. The polynomial r(x)=x+3 now maps 0 to 3 and 3 to 0. Its parameter belongs to B, and the result stays in B. More generally, polynomial operations whose parameters all come from B restrict to polynomial operations of B under the inherited signature.

This gives a precise repair to the vague claim that “subalgebras are closed under polynomials.” Specify which parameters are allowed. Closure under parameter-free terms follows from the subalgebra definition. Closure under arbitrary parameters from the whole ambient algebra is a stronger requirement and can fail immediately, as this two-element subset shows.

Contents · Previous chapter · Next chapter

One way to turn parameters into formal constants is to expand the signature. Add a nullary symbol for each chosen element and interpret it as that element. Terms in the expanded algebra then express the polynomial operations obtainable with those named parameters in the original algebra.

In our modulo-6 example, naming every carrier element makes all 36 unary affine functions term operations of the expanded algebra. The functions have not changed. Their classification relative to the language has changed because the available constant symbols have changed.

This expansion affects subalgebras. A subalgebra must contain the value of every named constant. If every element is named, any subalgebra must contain the whole carrier. The formerly valid subset {0,3} is no longer a subalgebra of that fully named expansion.

Congruences behave differently. A constant's compatibility condition relates its value to itself, already guaranteed by reflexivity. Adding constant symbols alone therefore does not remove congruences from a fixed algebra. The parity partition remains a congruence even after every residue receives a name.

The contrast is worth calculating rather than memorising. Subalgebra closure asks whether a selected subset contains the outputs of operations, including nullary outputs. Congruence compatibility asks whether corresponding related inputs produce related outputs; a nullary operation has no differing inputs and gives the same value on both sides. The two definitions explain why the same signature expansion has different consequences.

Contents · Previous chapter · Next chapter

The term “polynomial” can encourage a reader to write x² automatically. In the additive-group signature, multiplication is not a basic operation. The unary polynomial operations remain kx+c. A familiar expression from school algebra is not automatically available simply because the carrier elements are represented by numbers.

On residues modulo 6, the numerical squaring function has outputs 0,1,4,3,4,1. If it were affine, its values at 0 and 1 would force c=0 and k=1. That candidate is the identity function, but at input 2 it gives 2 rather than 4. Squaring is therefore not a polynomial operation of this additive algebra.

Expand the signature to the ring operations by including multiplication. Now x·x is a term, so the same squaring function becomes a term operation of the expanded algebra. The carrier remained unchanged, but a new operation enlarged what expressions could construct.

There is a subtle boundary: an unavailable expression can sometimes describe a function that is available through another term. On the additive group modulo 2, numerical squaring agrees with the identity function on both inputs. Although the multiplication symbol is absent from the additive signature, that two-entry function is induced by the permitted term x.

Thus “the written expression is not a term” and “the resulting function is not a term operation” are different claims. The first is a syntax check. The second asks whether any permitted term produces the same function. A correct impossibility proof must address the second question, as the complete affine classification did for modulo 6.

Contents · Previous chapter · Next chapter

Use the ordered carrier {0,1,2} with minimum, maximum and designated constants 0 and 2. These are lattice operations, not addition and multiplication. A unary polynomial may use fixed parameters from the chain, including the middle element 1. The resulting functions need a new derivation; the affine formula from the group example does not transfer.

Every unary polynomial on this chain is a clamp: choose a≤b and send x to min(max(x,a),b). Values below a rise to a, values above b fall to b, and values between them stay unchanged. The six choices of ordered bounds produce the output lists 000, 011, 012, 111, 112 and 222.

Why are there no more? Each variable or constant is a clamp. On a chain, taking the pointwise minimum or maximum of two clamps again gives a clamp: their lower and upper bounds are combined by the same minimum or maximum operation. Induction over a term with fixed parameters therefore stays within the six listed functions. Each listed clamp is explicitly constructible.

Without the parameter 1, the original terms can use only x and the two endpoint constants. Their unary functions are the constant 0, the identity and the constant 2. This follows by the same closure check on those three functions. Naming the middle element enlarges the unary term functions from three to six.

The lesson is not that every algebra has only a handful of polynomial operations. It is that classification follows the actual operations. A signature built from minimum and maximum imposes different restrictions from one built from group addition. Reusing an old formula because both carriers have numbered elements would erase the distinction universal algebra is designed to expose.

Contents · Previous chapter · Next chapter

Every minimum-and-maximum polynomial on the three-element chain is order-preserving. Yet not every order-preserving function is a polynomial operation of this lattice. Consider the output list 022: it sends 0 to 0 and both 1 and 2 to 2. The outputs never decrease, but the function is absent from the complete six-clamp list.

This gives a concrete warning about a necessary test. Passing monotonicity cannot establish polynomial representability. The full classification rules it out. Alternatively, its endpoint values force the bounds of any matching clamp to be a=0 and b=2, which would give the identity list 012 and therefore the wrong middle value.

For a different limitation, take the three-element carrier {0,1,2} with a unary operation f cycling 0→1→2→0 and no constants. Its only congruences are equality and the universal relation: identifying two distinct elements and applying the cycle forces all three together.

Every unary function preserves those two relations. But the unary polynomial operations of this cycle algebra are only the three powers of f and the three constant functions obtained by fixing parameters. A term can only apply f repeatedly to one variable or parameter; there is no operation combining branches. The table 001 is neither a cycle permutation nor constant, so it is congruence-preserving but not polynomial.

These examples show why proof direction matters. Polynomial operations preserve congruences; lattice polynomials are monotone. Neither statement supplies its converse in general. When a proposed test accepts too many functions, look for the additional construction restriction rather than announcing a false equivalence.

Contents · Previous chapter · Next chapter

First check the language. Write the carrier and the basic operations, including constants. Inspect the proposed expression for undeclared symbols. If fixed carrier elements appear, label them as parameters and state whether the question allows them. This avoids performing a correct calculation in the wrong signature.

Second check the function itself. State which inputs vary, list any fixed values and evaluate a small example. For a finite carrier, a complete table can reveal accidental equality between different expressions. Include the zero input when available, because it often exposes an offset that a few nonzero tests conceal.

Third check the property actually claimed. Term representability, polynomial representability, homomorphism preservation, congruence preservation and closure on a subalgebra are different requirements. Write the corresponding equation or inclusion. Do not let the word “preserves” stand alone without naming what is preserved.

Fourth check the scope of any generalisation. The classification kx+c was proved for an additive cyclic group. The clamp classification was proved for a specified chain lattice. Neither automatically covers nonabelian groups, arbitrary lattices or algebras with additional operations. A change in the signature can change the answer even when the carrier is unchanged.

For a practical notebook, retain the signature, parameter list, output table, claimed property and one verification or counterexample. That record is short enough to inspect and detailed enough to reconstruct. If another reader cannot determine whether a numeral names a constant symbol or a chosen parameter, improve the description before debating the calculation.

Contents · Previous chapter · Next chapter

Work in the additive group modulo 8 with signature +, negation and 0. Someone proposes p(x)=3x+2. Decide whether it is a unary term operation, a unary polynomial operation and a group endomorphism. Then test whether it maps the subalgebra B={0,4} into itself. Give a reason for each answer rather than treating the four labels as synonyms.

Next replace the parameter 2 by 4, keeping the coefficient 3 and the carrier unchanged. Recheck the four questions. Finally, name the element 4 with a new constant symbol and ask whether B remains a subalgebra of that expansion. Do the work before reading the explanation.

The first function is polynomial because it is a term with the fixed parameter 2. It is not a term operation of the original additive signature, because p(0)=2 rather than 0. The same value rules out a group endomorphism. Its outputs on B are p(0)=2 and p(4)=14=6 modulo 8, outside B.

For q(x)=3x+4, the function remains polynomial and still fails to be an original-signature term operation or group endomorphism, since q(0)=4. However, q(0)=4 and q(4)=16=0, so it maps B into itself. Its parameter belongs to B, consistent with the subalgebra reasoning earlier.

After naming 4, q becomes a term operation of the expanded algebra. B remains a subalgebra because it contains both named constants 0 and 4 and is closed under addition and negation. This task changes the modulus and parameter while keeping the reasoning framework. A successful answer tracks each change instead of importing the modulo-6 conclusions without checking them.

Contents · Previous chapter · Next chapter

Does every term operation count as a polynomial operation? Yes. Allowing parameters includes the case where none are needed. The inclusion may be strict, as the shift x↦x+1 shows in the additive-group example. In an algebra whose signature already names every element, the two collections coincide under the stated conventions.

Are constants always polynomial operations? For positive arity, a constant function can be represented by a term whose chosen parameter supplies the output and whose variable inputs are unused. Discussions of nullary operations and empty variable sets require careful conventions. Our examples use unary or higher-arity functions, so that technical boundary does not affect their conclusions.

Does a parameter have to be unknown? No. Here it is a chosen fixed element. The term “parameter” describes its role during evaluation, not uncertainty about its value. In another mathematical context, a family may allow that parameter to vary between functions. Within each selected polynomial operation, its chosen value is held fixed.

Can two different parameter choices produce the same function? They can. Our affine modulo-6 representation had unique coefficient pairs, and we proved that uniqueness using inputs 0 and 1. Other terms can ignore a parameter or combine parameters in ways that give identical outputs. Never assume that different written parameter lists imply different operations.

Why care about the distinction? It controls what constructions preserve. Original terms stay inside subalgebras and commute with homomorphisms in the appropriate evaluation sense. Fixed parameters must be included or transported when applying those statements. Making them explicit prevents a hidden extra element from invalidating an otherwise plausible structural argument.

Contents · Previous chapter · Next chapter

Begin by sorting a small collection of expressions in a declared signature. Use x+x, x+0, x+1 and x·x in the modulo-6 additive example. For each, explain whether the written expression is a term, whether a parameter is being supplied, and whether the described function might still have another permitted representation.

Next construct all six unary term tables and all 36 affine polynomial tables. You do not need to print a large wall of numbers. Group the functions by offset and show how changing the coefficient changes the successive differences. Use the classification proof to explain why the enumeration is complete.

Then test a function against several structures. The shift x↦x+1 is a useful candidate: it is bijective, is polynomial, preserves congruences and fails to preserve the group identity. On a selected subalgebra it may leave the subset. These simultaneous answers help break the habit of treating every useful map as a homomorphism.

Finally, move to the three-element chain and reconstruct the clamp argument without using addition. Identify the functions available before and after naming the middle element. Explain why the monotone table 022 is unavailable. This change of operations tests understanding more strongly than another set of numerical substitutions in the same affine formula.

The goal is a disciplined account of what an expression is allowed to use and what its resulting function actually preserves. Variables, designated constants and chosen parameters can occupy similar-looking places on a page while carrying different obligations. Once those roles are explicit, the calculations become easier to verify and their limits become easier to explain.

Contents · Previous chapter · Continue to related reading