A small zoo, not an infinite one
You have met functions one at a time across the last three modules: lines, parabolas, exponentials, logarithms, sine, cosine. It can feel like an endless bestiary.
It is not. Almost every function you will graph in this course is one of about ten basic shapes, the parent functions, possibly bent a little. Learn the ten on sight and you have the whole zoo.
Flip through the deck. For each card, notice the same five questions getting answered:
- End behavior: where does the curve go as runs far left and far right?
- Intercepts: where does it cross the axes?
- Asymptotes: is there a line the curve hugs but never touches?
- Periodicity: does the shape repeat?
- Symmetry: is it a mirror image across the -axis (even), or does it spin onto itself through the origin (odd)?
Those five questions are the vocabulary. You do not memorize graphs; you describe them.
Read the degree
A polynomial’s degree is its highest power of , and the degree controls its end behavior.
What is the degree of ?
Outside and inside
Here is the move that turns ten shapes into all of them.
Take any parent function . You can change its graph in exactly two places:
- Outside the function: . This happens after has fired. It moves the output.
- Inside the function: . This happens before fires. It moves the input.
Outside changes are vertical, because the output is the -value. Inside changes are horizontal, because the input is the -value. That single split, outside versus inside, is the whole grammar.
The full template is:
Four knobs: slides it sideways, slides it up and down, stretches it vertically, squeezes it horizontally.
Turn the knobs
Drive the four knobs yourself. The widget keeps the faint parent curve drawn so you can always see what moved.
Start with the parent . Drag the shift handle: the parabola slides, shape untouched. Drag the vertical stretch handle: it gets taller or, past zero, flips upside down. Drag the horizontal handle: it pinches inward.
Watch the formula and the plain-English step list rewrite themselves as you drag. That synchronization, picture and symbols moving together, is the thing to internalize. The formula is not a separate fact to memorize; it is a readout of what your hands just did.
Slide it sideways
Consider , the parent with its knobs set.
How many units to the right is this graph shifted?
Inside acts backward
One trap, and it catches everyone.
The graph of shifts right by 3, not left. The minus sign looks like it should pull the graph toward the negatives, but it does the opposite.
Here is why, with no hand-waving. The new graph at shows the value . So whatever the parent did at its origin now happens at . The origin moved right. Inside operations always run backward from what the sign suggests.
The same reversal hits the order of steps. Worked example: sketch .
- Inside, : shift right 3.
- Outside, the : stretch vertically by 2, then reflect across the -axis.
- Outside, the : shift up 5.
Outside operations apply in the normal order of operations. Inside operations apply in reverse. Get inside-versus-outside right and the rest follows.
Period of a squeezed wave
The parent repeats every . The inside multiplier in squeezes it horizontally by a factor of 2.
What is the period of ? (As a decimal.)
The wave was using this grammar all along
Switch the widget’s parent to .
Look at what the knobs become. The vertical stretch is the amplitude you met in m3. The horizontal squeeze is the frequency. The vertical shift is the midline offset.
Module 3 taught you amplitude, frequency, and offset as trigonometry vocabulary. They were never trig-specific. They are just with happening to be sine. One grammar, one widget, every parent function. The wave is not a special case; it is the same machine.
Find the midline
Consider . The amplitude is 3, the period is , and there is an outside shift.
The wave oscillates around the line what?
Where this shows up
A transformer is built from layers, and almost every layer is a transformation in exactly this sense. A learned weight matrix is, geometrically, a stretch and a reflection applied to its input vector. The positional encodings that tell the model where each token sits are shifted sine waves, with the and knobs turned.
You just learned the grammar those layers speak. Next we leave smooth curves for a moment and watch what happens when you add infinitely many numbers together.
Lesson complete