electronegativity
/ ee-lek-troh-neg-uh-TIV-ih-tee /
When two atoms share electrons in a bond, they rarely share fairly: one usually pulls harder. Electronegativity measures that pulling power, how strongly an atom, when bonded, tugs the shared electrons toward itself. Think of a tug-of-war over the bonding electrons; electronegativity is each atom's strength in that contest.
Linus Pauling put numbers to it on a scale where fluorine, the greediest of all, is about 4.0, oxygen about 3.4, carbon about 2.5, hydrogen about 2.2, and alkali metals like sodium and caesium sit near 0.8. The value rises across a period (left to right) and falls down a group. What really matters for bonding is the difference between two atoms: a large difference means one atom nearly wins outright and the electron is transferred (ionic), a tiny difference means a fair share (covalent between like atoms), and a moderate difference gives a lopsided share (a polar covalent bond).
This single number is the best quick predictor of bond type. A common rule of thumb: a difference above about 1.7 to 2.0 leans ionic, below that leans covalent, and near zero between two metals gives metallic bonding. Sodium (0.9) and chlorine (3.0) differ by 2.1, strongly ionic, hence salt. Carbon and hydrogen differ by only 0.35, nearly nonpolar covalent, the backbone of organic chemistry. Electronegativity difference is the dial that sweeps from one bond type to another, which is why it reappears under mixed bonding.
In water, oxygen (3.4) badly out-pulls hydrogen (2.2), so each O-H bond is polar: oxygen carries a slight negative charge, the hydrogens a slight positive. That built-in lopsidedness is what makes water a good solvent and gives rise to the hydrogen bond that holds ice together.
The difference in electronegativity, not the raw values, predicts the bond type.
Electronegativity is a derived, model-dependent concept, not a directly measured atomic property; different scales (Pauling, Mulliken, Allred-Rochow) give slightly different numbers. The 1.7 ionic-covalent cutoff is a rough guide with many exceptions, not a hard boundary.