The Central Idea
Consider the sequence
\[ 1,\ 4,\ 9,\ 16,\ 25,\ldots \]The first differences are
\[ 3,\ 5,\ 7,\ 9,\ldots \]These are not constant, so the sequence is not arithmetic.
Now take the differences again:
\[ 5-3=2, \] \[ 7-5=2, \] \[ 9-7=2. \]The second differences are constant:
\[ \boxed{2,\ 2,\ 2,\ldots} \]This constant second difference reveals the quadratic rule
\[ \boxed{a_n=n^2}. \]What Is a Finite Difference?
A finite difference measures the change between terms at discrete indices.
The first difference is
\[ \boxed{\Delta a_n=a_{n+1}-a_n}. \]Here, the symbol
\[ \Delta \]represents a change or difference.
Repeatedly applying the difference operation produces second, third, and higher differences.
First Differences
For the sequence
\[ 4,\ 9,\ 14,\ 19,\ 24,\ldots \]the first differences are
\[ 9-4=5, \] \[ 14-9=5, \] \[ 19-14=5. \]Therefore,
\[ \boxed{\Delta a_n=5}. \]Constant first differences identify an arithmetic sequence.
Second Differences
Suppose the first differences are not constant.
For
\[ 2,\ 6,\ 12,\ 20,\ 30,\ldots \]the first differences are
\[ 4,\ 6,\ 8,\ 10,\ldots \]Take differences again:
\[ 6-4=2, \] \[ 8-6=2, \] \[ 10-8=2. \]Thus the second differences are
\[ \boxed{2,\ 2,\ 2,\ldots}. \]Constant second differences indicate quadratic structure.
Third Differences
Cubic sequences generally do not have constant first or second differences, but they do have constant third differences.
Consider
\[ 1,\ 8,\ 27,\ 64,\ 125,\ldots \]generated by
\[ a_n=n^3. \]The first differences are
\[ 7,\ 19,\ 37,\ 61,\ldots \]the second differences are
\[ 12,\ 18,\ 24,\ldots \]and the third differences are
\[ \boxed{6,\ 6,\ldots}. \]The General Polynomial Pattern
If a sequence is generated by a polynomial of degree \(k\), then its \(k\)th finite differences are constant, provided the indices increase by equal steps.
Thus:
| Polynomial Degree | Constant Difference Level |
|---|---|
| 0 | Terms themselves are constant |
| 1 | First differences |
| 2 | Second differences |
| 3 | Third differences |
| \(k\) | \(k\)th differences |
A Difference Table
A difference table organizes a sequence and its repeated differences.
For the square numbers,
\[ 1,\ 4,\ 9,\ 16,\ 25, \]the structure is
| Terms | First Differences | Second Differences |
|---|---|---|
| \(1\) | \(3\) | \(2\) |
| \(4\) | \(5\) | \(2\) |
| \(9\) | \(7\) | \(2\) |
| \(16\) | \(9\) | |
| \(25\) |
The constant second-difference column reveals the quadratic pattern.
Why Each Difference Row Is Shorter
A difference requires two adjacent values.
If the original row contains five terms, then only four first differences can be formed.
Those four values produce only three second differences.
In general, each successive difference row contains one fewer value than the row above it.
Arithmetic Sequences Revisited
Arithmetic sequences are the degree-one case of polynomial sequences.
For
\[ a_n=4n+2, \]the terms are
\[ 6,\ 10,\ 14,\ 18,\ldots \]and the first differences are
\[ \boxed{4,\ 4,\ 4,\ldots}. \]The constant first difference equals the coefficient of \(n\).
Quadratic Sequences
A quadratic sequence has an explicit rule of the form
\[ a_n=An^2+Bn+C \]with
\[ A\neq0. \]Such a sequence has constant second differences.
Even when the quadratic rule is not immediately visible, the difference table exposes its degree.
Example: Recognizing a Quadratic Sequence
Consider
\[ 3,\ 8,\ 15,\ 24,\ 35,\ldots \]The first differences are
\[ 5,\ 7,\ 9,\ 11,\ldots \]and the second differences are
\[ 2,\ 2,\ 2,\ldots \]Therefore, the sequence has quadratic structure.
In fact,
\[ \boxed{a_n=n^2+2n}. \]The Leading Coefficient and Second Differences
For a quadratic sequence
\[ a_n=An^2+Bn+C, \]the constant second difference is
\[ \boxed{2A}. \]Therefore, if the constant second difference is \(6\), then
\[ 2A=6 \]and
\[ A=3. \]This gives us an efficient starting point for reconstructing a quadratic rule.
Deriving the Quadratic Difference Pattern
Suppose
\[ a_n=An^2+Bn+C. \]Then
\[ a_{n+1} = A(n+1)^2+B(n+1)+C. \]Expand:
\[ a_{n+1} = An^2+2An+A+Bn+B+C. \]Subtract \(a_n\):
\[ \Delta a_n = 2An+A+B. \]The first differences are linear in \(n\).
Taking another difference gives
\[ \boxed{\Delta^2a_n=2A}. \]Example: Recovering a Quadratic Formula
Consider
\[ 4,\ 11,\ 22,\ 37,\ 56,\ldots \]The first differences are
\[ 7,\ 11,\ 15,\ 19, \]and the second differences are
\[ 4,\ 4,\ 4. \]Since
\[ 2A=4, \]we have
\[ A=2. \]So write
\[ a_n=2n^2+Bn+C. \]Solving for the Remaining Coefficients
Continue with
\[ a_n=2n^2+Bn+C. \]Since \(a_1=4\),
\[ 2+B+C=4, \]so
\[ B+C=2. \]Since \(a_2=11\),
\[ 8+2B+C=11, \]so
\[ 2B+C=3. \]Subtracting the equations gives
\[ B=1. \]Therefore,
\[ C=1. \]The rule is
\[ \boxed{a_n=2n^2+n+1}. \]Checking the Quadratic Rule
Test
\[ a_n=2n^2+n+1 \]at \(n=3\):
\[ a_3=2(3)^2+3+1 \] \[ =18+3+1 \] \[ =22. \]This matches the third term of
\[ 4,\ 11,\ 22,\ 37,\ 56,\ldots \]so the recovered rule is consistent with the sequence.
Cubic Sequences
A cubic sequence has an explicit rule of the form
\[ a_n=An^3+Bn^2+Cn+D \]with
\[ A\neq0. \]Its third finite differences are constant.
The second differences are generally arithmetic, and the first differences are generally quadratic.
Example: Recognizing a Cubic Sequence
Consider
\[ 1,\ 8,\ 27,\ 64,\ 125. \]First differences:
\[ 7,\ 19,\ 37,\ 61. \]Second differences:
\[ 12,\ 18,\ 24. \]Third differences:
\[ 6,\ 6. \]The constant third difference identifies cubic structure.
Here,
\[ \boxed{a_n=n^3}. \]The Leading Coefficient and Higher Differences
For a polynomial sequence of degree \(k\),
\[ a_n=An^k+\text{lower-degree terms}, \]the constant \(k\)th difference is
\[ \boxed{k!A}. \]For a quadratic,
\[ 2!A=2A. \]For a cubic,
\[ 3!A=6A. \]For a quartic,
\[ 4!A=24A. \]Why Factorials Appear
Each difference operation lowers polynomial degree by one.
The leading term
\[ An^k \]eventually becomes a constant after \(k\) successive differences.
The repeated coefficient changes accumulate the factor
\[ k(k-1)(k-2)\cdots2\cdot1 \]which is
\[ \boxed{k!}. \]Finite Differences as a Discrete Analogue of Derivatives
Finite differences play a role in discrete mathematics similar to the role derivatives play in calculus.
Differentiation lowers polynomial degree:
\[ n^3\longrightarrow3n^2\longrightarrow6n\longrightarrow6. \]Finite differencing also lowers polynomial degree:
\[ \text{cubic sequence} \longrightarrow \text{quadratic differences} \longrightarrow \text{linear differences} \longrightarrow \text{constant differences}. \]The operations are not identical, but the structural analogy is extremely useful.
The Difference Operator
The first finite-difference operator is written
\[ \Delta. \]Thus,
\[ \Delta a_n=a_{n+1}-a_n. \]Applying the operator again gives
\[ \Delta^2a_n = \Delta(\Delta a_n). \]A third application gives
\[ \Delta^3a_n. \]The exponent on \(\Delta\) counts how many times the difference operation has been applied.
Second Difference in Expanded Form
Starting from
\[ \Delta a_n=a_{n+1}-a_n, \]we have
\[ \Delta^2a_n = \Delta a_{n+1}-\Delta a_n. \]Therefore,
\[ \Delta^2a_n = (a_{n+2}-a_{n+1})-(a_{n+1}-a_n). \]Simplifying,
\[ \boxed{ \Delta^2a_n = a_{n+2}-2a_{n+1}+a_n }. \]Predicting the Next Term with Differences
Difference tables can be used to continue a polynomial pattern.
Consider
\[ 2,\ 6,\ 12,\ 20,\ 30. \]First differences:
\[ 4,\ 6,\ 8,\ 10. \]Second differences:
\[ 2,\ 2,\ 2. \]Continue the constant second difference with another \(2\).
The next first difference is
\[ 10+2=12. \]Therefore, the next term is
\[ 30+12=\boxed{42}. \]Extending a Cubic Pattern
Consider
\[ 1,\ 8,\ 27,\ 64. \]First differences:
\[ 7,\ 19,\ 37. \]Second differences:
\[ 12,\ 18. \]Third difference:
\[ 6. \]Continue the third difference with another \(6\).
The next second difference is
\[ 18+6=24. \]The next first difference is
\[ 37+24=61. \]The next term is
\[ 64+61=\boxed{125}. \]Working Backward with Differences
A difference table can also reconstruct earlier terms.
Suppose
\[ \square,\ 9,\ 16,\ 25 \]follows the square-number pattern.
The first differences ending at these terms are
\[ 7,\ 9. \]Since the second difference is \(2\), the preceding first difference must be
\[ 7-2=5. \]Therefore, the missing term is
\[ 9-5=\boxed{4}. \]Difference Tables and Interpolation
A finite set of equally spaced data values can always be matched by a polynomial of sufficiently low degree.
Difference tables help reveal that polynomial structure.
If \(m+1\) data values are given at equally spaced indices, there is a polynomial of degree at most \(m\) that passes through them.
This idea forms part of polynomial interpolation, which will appear again in later mathematics.
A Pattern Can Fit Several Rules
A short list of terms does not uniquely determine what comes next unless additional assumptions are made.
For example,
\[ 1,\ 4,\ 9,\ 16 \]strongly suggests
\[ a_n=n^2. \]But another higher-degree polynomial can be constructed that agrees with these first four values and then differs later.
Difference-table continuation assumes that the observed polynomial pattern is intended to continue.
Why This Does Not Contradict Pattern Recognition
Pattern recognition is still mathematically valuable.
The important distinction is between
- identifying a simple rule consistent with the data; and
- proving that the data logically force that rule.
In exercises about finite differences, we normally assume that the simplest polynomial structure revealed by the difference table is intended.
Polynomial Sequences Are Discrete Polynomial Functions
A polynomial sequence is simply a polynomial function evaluated at discrete index values.
For example,
\[ f(x)=x^2+1 \]is a continuous polynomial function when \(x\) ranges over the real numbers.
The sequence
\[ a_n=n^2+1,\qquad n=1,2,3,\ldots \]uses the same algebraic rule but only at the allowed integer indices.
Sequence Tables and Polynomial Graphs
The sequence
\[ a_n=n^2 \]produces the points
\[ (1,1),\ (2,4),\ (3,9),\ (4,16),\ldots \]These points lie on the parabola
\[ y=x^2. \]The polynomial curve provides a continuous extension, while the sequence itself remains discrete.
Equal Index Spacing Matters
Standard finite-difference tests assume that the input indices are equally spaced.
For example, comparing
\[ a_1,\ a_2,\ a_3,\ a_4 \]uses index steps of \(1\).
If the inputs are instead
\[ 1,\ 2,\ 4,\ 7, \]ordinary differences of the outputs cannot be interpreted in the same way because the horizontal step sizes differ.
Finite Differences with Another Constant Step
The index spacing does not have to be \(1\), but it must be constant for the usual difference-table structure.
For example, evaluating a quadratic at
\[ x=0,\ 2,\ 4,\ 6 \]still produces constant second differences because each input step is consistently \(2\).
The numerical constant will depend on the step size.
Example: A Quadratic with Step Size Two
Let
\[ f(x)=x^2. \]Evaluate at
\[ x=0,\ 2,\ 4,\ 6. \]The values are
\[ 0,\ 4,\ 16,\ 36. \]First differences:
\[ 4,\ 12,\ 20. \]Second differences:
\[ \boxed{8,\ 8}. \]The second differences remain constant, although they are no longer \(2\) because the input spacing has changed.
Constant Differences Do Not Mean Constant Terms
If first differences are constant, the terms themselves generally change.
For example,
\[ 3,\ 8,\ 13,\ 18,\ldots \]has constant first difference
\[ 5. \]The sequence is not constant; its rate of change is constant.
Higher Differences Eventually Become Zero
Once a polynomial sequence reaches its constant difference level, one more difference produces zeros.
For a quadratic sequence,
\[ \Delta^2a_n=\text{constant}. \]Therefore,
\[ \Delta^3a_n=0. \]More generally, a degree-\(k\) polynomial sequence satisfies
\[ \boxed{\Delta^{k+1}a_n=0}. \]Example: Squares Reach Zero at the Third Difference
Start with
\[ 1,\ 4,\ 9,\ 16,\ 25. \]First differences:
\[ 3,\ 5,\ 7,\ 9. \]Second differences:
\[ 2,\ 2,\ 2. \]Third differences:
\[ \boxed{0,\ 0}. \]This is consistent with a degree-two polynomial.
Newton's Forward-Difference View
Finite differences can do more than classify polynomial degree.
They can also build a polynomial sequence from its initial value and difference data.
For a sequence indexed from \(0\), one form is
\[ a_n = a_0 + \binom{n}{1}\Delta a_0 + \binom{n}{2}\Delta^2a_0 + \binom{n}{3}\Delta^3a_0 +\cdots. \]For a polynomial sequence, this expression eventually terminates because sufficiently high differences become zero.
A Gentle Example of the Forward-Difference Formula
For the square numbers beginning at \(n=0\),
\[ 0,\ 1,\ 4,\ 9,\ldots \]we have
\[ a_0=0, \] \[ \Delta a_0=1, \]and
\[ \Delta^2a_0=2. \]Therefore,
\[ a_n = 0 + \binom{n}{1}(1) + \binom{n}{2}(2). \]Since
\[ \binom{n}{1}=n \]and
\[ 2\binom{n}{2}=n(n-1), \]we obtain
\[ a_n=n+n(n-1) \] \[ =\boxed{n^2}. \]Why Binomial Coefficients Appear
Repeated finite differences interact naturally with the polynomial basis
\[ 1,\quad \binom{n}{1},\quad \binom{n}{2},\quad \binom{n}{3},\ldots \]because differencing these expressions lowers their index in a particularly simple way.
This connection ties sequence patterns to combinatorics, Pascal's triangle, interpolation, and discrete calculus.
Difference Tables as Discrete Structure
A difference table does not merely provide a trick for guessing the next number.
It exposes layers of algebraic structure:
- the original sequence gives the values;
- first differences describe discrete change;
- second differences describe change in that change;
- higher differences continue the same idea.
This layered view is one of the basic organizing ideas of discrete analysis.
Common Error: Stopping at Nonconstant First Differences
Suppose a student examines
\[ 1,\ 4,\ 9,\ 16,\ldots \]and notes that
\[ 3,\ 5,\ 7,\ldots \]are not constant.
Concluding that there is no simple polynomial pattern would be premature.
Take another difference:
\[ 2,\ 2,\ldots \]The constant second difference reveals quadratic structure.
Common Error: Assuming Constant Second Differences Mean Cubic
Constant second differences indicate a polynomial of degree at most \(2\), not degree \(3\).
The pattern is:
\[ \text{constant first differences}\longrightarrow\text{linear}, \] \[ \text{constant second differences}\longrightarrow\text{quadratic}, \] \[ \text{constant third differences}\longrightarrow\text{cubic}. \]Common Error: Forgetting to Use Consecutive Terms
Standard first differences compare adjacent sequence terms:
\[ a_{n+1}-a_n. \]Skipping irregularly between terms changes the step size and can obscure the pattern.
Difference tables should normally be built from equally spaced indices.
Common Error: Treating Difference Tables as Proof of a Unique Future
A finite list of values can be matched by many possible rules.
A difference table identifies a polynomial rule consistent with the given data, usually the simplest one implied by the constant-difference pattern.
It does not prove that no other rule could agree with the same initial values and then behave differently.
Common Error: Ignoring the Indexing Convention
The formulas
\[ a_n=n^2 \]for \(n\geq1\) and
\[ b_n=(n+1)^2 \]for \(n\geq0\) generate the same displayed values:
\[ 1,\ 4,\ 9,\ 16,\ldots \]but they attach different indices to those values.
Finite differences depend on the indexed data, so the indexing convention remains part of the definition.
A Reliable Procedure for Building a Difference Table
- Write the sequence terms in order.
- Subtract each term from the next to form first differences.
- If the first differences are not constant, difference them again.
- Continue until a constant row appears or the available data run out.
- Count the number of difference levels required.
- Use that level to identify the likely polynomial degree.
- Extend the constant row if the pattern is assumed to continue.
- Work upward to predict additional terms when needed.
A Reliable Procedure for Recovering a Quadratic Rule
- Build the difference table.
- Verify that the second differences are constant.
- Set \(2A\) equal to the constant second difference.
- Solve for \(A\).
- Write \(a_n=An^2+Bn+C\).
- Substitute two known sequence terms.
- Solve the resulting system for \(B\) and \(C\).
- Check the formula against another known term.
Example: Full Difference-Table Workflow
Consider
\[ 5,\ 12,\ 23,\ 38,\ 57. \]First differences:
\[ 7,\ 11,\ 15,\ 19. \]Second differences:
\[ 4,\ 4,\ 4. \]Therefore, the sequence is quadratic.
The next first difference is
\[ 19+4=23. \]The next term is
\[ 57+23=\boxed{80}. \]Example: Recovering the Formula
Continue with
\[ 5,\ 12,\ 23,\ 38,\ 57,\ldots \]The constant second difference is \(4\), so
\[ 2A=4 \]and
\[ A=2. \]Write
\[ a_n=2n^2+Bn+C. \]Using \(a_1=5\),
\[ 2+B+C=5, \]so
\[ B+C=3. \]Using \(a_2=12\),
\[ 8+2B+C=12, \]so
\[ 2B+C=4. \]Subtract:
\[ B=1. \]Then
\[ C=2. \]Therefore,
\[ \boxed{a_n=2n^2+n+2}. \]Practice 1: First Differences
Find the first differences of
\[ 7,\ 12,\ 17,\ 22,\ 27. \]What type of sequence is it?
Practice 2: Second Differences
Build the first and second differences for
\[ 2,\ 7,\ 14,\ 23,\ 34. \]What polynomial degree is suggested?
Practice 3: Predict the Next Term
Assuming a quadratic pattern continues, find the next term:
\[ 3,\ 8,\ 15,\ 24,\ 35,\ldots \]Practice 4: Recognize a Cubic Sequence
Build successive differences for
\[ 1,\ 8,\ 27,\ 64,\ 125. \]At which difference level do the values become constant?
Practice 5: Find the Leading Coefficient
A quadratic sequence has constant second difference
\[ 10. \]Find the coefficient \(A\) in
\[ a_n=An^2+Bn+C. \]Practice 6: Recover a Quadratic Formula
Find an explicit formula for
\[ 3,\ 10,\ 21,\ 36,\ 55,\ldots \]assuming indexing begins at \(1\).
Practice 7: Work Backward
The sequence
\[ \square,\ 16,\ 25,\ 36 \]has constant second difference \(2\).
Find the missing term.
Practice 8: Determine Whether the Sequence Is Quadratic
Determine whether
\[ 2,\ 6,\ 18,\ 54,\ldots \]has constant second differences.
Practice 9: Higher-Degree Structure
A polynomial sequence has constant fourth differences.
What is the highest possible degree of the simplest polynomial generating the sequence?
Practice 10: Identify the Error
A student says that because
\[ 1,\ 4,\ 9,\ 16 \]has constant second differences, the next term is logically forced to be \(25\).
Explain what qualification is needed.
Practice Answers
Practice 1
The first differences are
\[ 5,\ 5,\ 5,\ 5. \]Therefore, the sequence is
\[ \boxed{\text{arithmetic}} \]and is generated by a degree-one polynomial.
Practice 2
First differences:
\[ 5,\ 7,\ 9,\ 11. \]Second differences:
\[ 2,\ 2,\ 2. \]Therefore, the pattern suggests a
\[ \boxed{\text{quadratic}} \]sequence.
Practice 3
First differences:
\[ 5,\ 7,\ 9,\ 11. \]Second differences:
\[ 2,\ 2,\ 2. \]Continue with another second difference of \(2\):
\[ 11+2=13. \]Therefore, the next term is
\[ 35+13=\boxed{48}. \]Practice 4
First differences:
\[ 7,\ 19,\ 37,\ 61. \]Second differences:
\[ 12,\ 18,\ 24. \]Third differences:
\[ 6,\ 6. \]Thus the values become constant at the
\[ \boxed{\text{third difference}}. \]Practice 5
For a quadratic sequence,
\[ \Delta^2a_n=2A. \]Therefore,
\[ 2A=10, \]so
\[ \boxed{A=5}. \]Practice 6
First differences:
\[ 7,\ 11,\ 15,\ 19. \]Second differences:
\[ 4,\ 4,\ 4. \]Therefore,
\[ 2A=4 \]and
\[ A=2. \]Write
\[ a_n=2n^2+Bn+C. \]From \(a_1=3\),
\[ B+C=1. \]From \(a_2=10\),
\[ 2B+C=2. \]Therefore,
\[ B=1,\qquad C=0. \]Hence,
\[ \boxed{a_n=2n^2+n}. \]Practice 7
The known terms
\[ 16,\ 25,\ 36 \]have first differences
\[ 9,\ 11. \]With constant second difference \(2\), the preceding first difference is
\[ 9-2=7. \]Therefore, the missing term is
\[ 16-7=\boxed{9}. \]Practice 8
First differences:
\[ 4,\ 12,\ 36. \]Second differences:
\[ 8,\ 24. \]These are not constant.
Therefore, the sequence is
\[ \boxed{\text{not quadratic by the finite-difference test}}. \]Practice 9
Constant fourth differences indicate a polynomial of degree at most
\[ \boxed{4}. \]If the fourth differences are nonzero, the polynomial has degree exactly \(4\).
Practice 10
If we assume the sequence continues according to the quadratic pattern revealed by the constant second differences, then the next term is \(25\).
However, a finite list of values does not uniquely determine every possible continuation.
Therefore, the correct statement is:
\[ \boxed{ \text{under the quadratic-pattern assumption, the next term is }25 }. \]What to Remember
Finite differences measure discrete change:
\[ \boxed{\Delta a_n=a_{n+1}-a_n}. \]For polynomial sequences with equally spaced indices:
- constant first differences indicate linear structure;
- constant second differences indicate quadratic structure;
- constant third differences indicate cubic structure;
- constant \(k\)th differences indicate degree at most \(k\);
- a degree-\(k\) polynomial has zero \((k+1)\)th differences;
- the constant \(k\)th difference equals \(k!\) times the leading coefficient.
The central structural pattern is
\[ \boxed{ \text{polynomial degree} \longleftrightarrow \text{level of constant finite difference} }. \]Difference tables therefore provide both a recognition tool and a bridge into interpolation, combinatorics, algorithms, and discrete calculus.
Where We Go Next
Polynomial sequences can often be understood by examining how their differences evolve. Other sequences are defined most naturally by rules that refer directly to earlier terms rather than by a simple explicit polynomial formula.
The next lesson, Fibonacci-Type Recurrences, develops sequences in which later terms depend on multiple preceding terms and explores how recursive structure can generate rich algebraic patterns.